Рубрика: AI

With AI becoming mainstream, demand for data scientists has climbed sharply across experience levels over the past year, extending beyond Tier 1 hubs and into Tier 2 cities.

A recent TeamLease report points to demand for nearly 45,000 AI-related roles, led by data scientists and machine learning engineers. LinkedIn data further underscores this momentum, ranking data science as the fastest-growing job category globally, with more than 11 million roles projected by 2026.

Bengaluru continues to be the undisputed leader across the talent spectrum, recording the highest share of demand at junior (34.60%), mid-level (40.48%), and senior (36.54%) roles, with notable growth across all three bands, according to the Randstad Digital Technology Skills Insights Report for India.

However, Hyderabad maintains a solid second position nationwide, demonstrating consistent demand across junior (15.89%), mid-level (17.25%), and senior (14.72%) roles.

Data scientist positions remain heavily concentrated in Tier 1 metropolitan hubs, led by Bengaluru (38.02%), followed by Hyderabad (16.08%), Pune (11.74%), Mumbai (11.46%), and Delhi-NCR (9.34%).

“Hyderabad has been steadily growing into a strategic location for many new GCCs to set up their primary centres in India. Many of these GCCs are hiring for AI, applied analytics, and Data Engineering,” Neeti Sharma, CEO, TeamLease Digital told AIM.

Proptech firm StarRez wants an early-mover advantage by tapping into the city’s data science talent pool. The company told AIM it is in the early stages of setting up a dedicated data science team in Hyderabad. The US-headquartered student housing management platform is making a deliberate push into India, with Hyderabad as an innovation hub for end-to-end product development, and a strong focus on AI, data science, and automation-led outcomes for universities globally.

The company hired its first employee in India in April this year and currently has a team of around 40 people in Hyderabad. Over the next 18 months, StarRez plans to scale this number to nearly 200.

“To go with Hyderabad was very deliberate,” Baumgartner stated. “We’re trying to build a true innovation hub for end-to-end product delivery—from product management, engineering platforms and architecture, all the way through to customer impact.”

He mentioned that Hyderabad stood out not only for its deep engineering talent but also for its strong product mindset, a combination he found to be rare.

“The talent ecosystem here combines deep engineering skill with product thinking,” he said. “What really stood out to me, after visiting multiple times and interviewing candidates, is how well the local mindset aligns with StarRez’s culture—very practical about innovation, ownership, and long-term value creation.”

A Nascent but Strategic Opportunity

Alongside AI, StarRez is also laying the foundation for a dedicated data science team in Hyderabad.

“We’re building a data science team in Hyderabad as well,” Baumgartner said. “We don’t have any other data capabilities yet at StarRez, and that’s a huge opportunity for us that’s currently nascent.”

The goal is to move beyond reactive insights and towards predictive and personalised outcomes through what the company calls an “intelligent resident ecosystem.” This includes analysing millions of data points—from payment behaviour to social engagement—to identify early signals of student distress and prompt timely human intervention.

The company operates under global compliance frameworks, including SOC 2 Type II, and uses fine-grained data residency controls across its Azure-based infrastructure.

StarRez partners closely with Microsoft and uses Microsoft OpenAI models, alongside orchestration technologies such as Semantic Kernel and the upcoming Microsoft Agent Framework.

From Full-Stack Delivery to AI-Led Innovation

StarRez is building the Hyderabad innovation hub as a full-stack capability centre, contributing across product development, engineering, platforms, quality, AI, and data science.

“GenAI, agentic AI, and data science are going to be major focus areas for us,” Baumgartner noted.

The company, which serves over 1,200 institutions globally and more than 3 million residents, is increasingly embedding AI into its platform to help university housing teams reduce operational burden and focus on meaningful human interaction with students.

One of the early AI implementations began with simple writing tools powered by large language models, aimed at streamlining routine communication tasks for housing administrators. This evolved into RES 360 Intelligence, a copilot-style interface that allows users to interact with student and housing data through natural language.

“That copilot experience is really our beachhead,” the CTPO explained. “It sets us up to launch agentic workflows.”

StarRez is now working towards task-based agentic AI workflows, where individual AI agents handle specific tasks such as room swaps, check-ins or maintenance requests. Over time, these will be orchestrated into role-based agents that mirror real-world housing roles.

A standout example is StarRez’s AI-powered inspection feature, which won the Vista AI Hackathon Award. The tool uses computer vision to inspect rooms via mobile devices, reducing inspection time from 11 minutes to 30 seconds per room.

“For a university like Columbia with 5,000 beds, that translates to nearly 700 workdays saved per year,” Baumgartner noted. “We’re very intentional about matching technology to measurable customer impact.”

Scaling Talent and Building Brand Presence

For now, StarRez’s hiring in India is focused on experienced professionals across software engineering, AI, data science, platform reliability, product, and design. Fresher hiring will follow once foundational leadership and delivery teams are firmly in place.

While brand awareness is a challenge for newer entrants, StarRez is actively investing in building its presence in Hyderabad through university partnerships, hackathons, meetups and ecosystem engagement, supported by its India partner Zinnov.

“Branding is a challenge for every company that comes in without existing brand equity,” Baumgartner said. “But there are strategies to move the needle quickly, and we’re leaning into them.”

“This team is already shipping features,” he added. “But what excites us is what’s next, connecting AI, data, automation and human insight to create environments that help students thrive, not just live.”

The post GCCs Still Can’t Get Enough Of Data Scientist Talent, And StarRez Knows That appeared first on Analytics India Magazine.

When a deepfake Jensen Huang livestream promoting cryptocurrency pulled in 100,000 viewers, eight times NVIDIA’s actual GPU Technology Conference audience, it exposed a failure that was more social than technical.

A scam channel labelled “NVIDIA Live” used a QR code to funnel victims into sending cryptocurrency.

The render didn’t have to be flawless, but convincing enough to bypass scepticism.

“In information security, there is no such thing as being one step ahead of the cybercriminals. That sounds very beautiful in some studies — but it’s simply wishful thinking,” Miguel Fornés, security governance and compliance manager at Surfshark, a cybersecurity firm, said in an interaction with AIM.

Surfshark’s analysis places deepfake-related financial losses at $1.5 billion in 2025 alone. Fornés expects that number to grow as generative models improve.

The more critical admission is that defenders must be active at all times, while attackers need only one opportunity.

Two years ago, viral deepfakes were visibly flawed. “Just 24 months later, everything changed dramatically,” Fornés says.

From OpenAI’s Sora 2 to Google’s Veo 3.1, and even open-weighted text-to-video models from Chinese AI labs such as Tencent, have advanced rapidly.

“There are hundreds of millions of dollars right now, every day, being thrown at the most important AI companies to reach dominance, to reach superiority,” said Fornés.

But once that superiority is achieved by these companies — it is offered for nominal prices when packaged into a product.
“Previously, depending on quality, producing a one-minute deepfake video was estimated to cost between $300 and $20,000 and required professional skills,” stated Surfshark’s study.

But these costs have drastically reduced, widening the defender’s disadvantage.

Besides, traditional technical tells are eroding quickly. Looking out for signs of unusually clean audio, recycled footage, physics errors and visual artefacts may help now, but they won’t be reliable for long.

Can We Even Identify Deepfakes?

Talking about the limits of visual detection, Fornés said, “Whatever I tell you now, in one month, it won’t be valid. It’s as simple as that.”

Even so, he outlined a few signals users have traditionally relied on that are already eroding.
One of the clearest tells, he explained, used to be audio quality.

Real recordings carry the messiness of everyday life. A background chatter, a TV humming, appliances whirring, someone walking past an office desk. Deepfakes often strip all this away, resulting in unnaturally clean audio that should prompt suspicion.
Another pattern he pointed out was the reuse of existing footage. Many malicious clips aren’t fully generated from scratch.
Attackers “take some existing video” and simply modify the face or key characteristics. That means reverse-image searches sometimes catch the original material beneath the manipulation.

But visual inconsistencies, he warned, are disappearing just as quickly.

He recalled a deepfake video he saw circulating on Facebook — supposedly showing baggage handlers aggressively throwing suitcases.

What caught his eye wasn’t the outrage but the physics. One bag is thrown into a compartment and then inexplicably bounces back out despite a door blocking the path.

“This is what is called illogical kinetics,” he said, the kind of impossible movement that signals fakery.

The problem is that these tells are living on borrowed time. These illogical kinetics “will be fixed” in a few months, Fornés said.
The jerky motions, the uncanny rebounds, the unnatural steadiness in a synthetic face, all of it is being ironed out by rapidly improving models. The list of practical, user-visible indicators is evaporating.

That’s why he rejects the idea that this is a purely technical problem. The core vulnerability isn’t the pixels but the human mind.

“What happens with cybercrime is that it’s a psychological issue… they are pulling emotional clickers, and they are trying to elicit an emotional response,” said Fornés.

He recommends behaviourally focused defences, as in when media elicits a strong emotional reaction, especially around politics, unrest, or requests for money or action, pause and verify before acting.
Experts now recommend families decide a secret code. When a voice that sounds precisely like your kin’s, requesting urgent financial help, you ask for the code.

It’s a profoundly dystopian adaptation, one where verifying loved ones requires prearranged passwords, and trust itself has become a liability.

The C2PA Hope

Fornés highlights content provenance as the strongest systemic mitigation available today, pointing to the C2PA standard.
C2PA embeds “Content Credentials” directly into media—cryptographically sealed metadata that records when the asset was created, by whom or what system, and how it has been altered over time.
It is a joint effort that merges Adobe’s Content Authenticity Initiative with Project Origin, the BBC–CBC–New York Times–Microsoft consortium, and now includes major stakeholders across tech and media such as Adobe, Amazon, BBC, Google, Intel, Meta, Microsoft, OpenAI, Publicis Groupe, Sony and Truepic.

Each update produces a new signed entry, and the manifest is bound to the file’s pixel or byte structure. If the media is changed without updating the manifest, verification breaks.
In effect, the file carries a tamper-evident trail of its origin and transformations, allowing any newly generated image or video to be explicitly identifiable.

Only the other way around is viable: Attach cryptographic proof that the image I am looking at has been taken with a real camera and not altered.
This is what @C2PA_org is solving, already implemented and supported by many.
Think "source provenance", but for media.

— Steren (@steren) December 1, 2025

Fighting Slop

Recently, researchers at Mila, Québec’s AI institute, introduced OpenFake, released a new dataset to train the next generation of deepfake detection tools.

The problem, they argue, is not that detection tools are failing in isolation.

“Most of these issues were resolved in the latest models,” the researchers noted, leaving detectors trained on legacy data “almost incapable” of flagging modern deepfakes in the wild.

OpenFake dataset pairs three million real images with one million synthetic counterparts generated using current state-of-the-art models, including OpenAI’s GPT-Image-1, Google’s Imagen-4, and Flux 1.1 Pro.

“While humans can easily get fooled by visual realism, our method can spot tiny artefacts produced when an AI tool generates or upscales images,” the paper explains.

When Mila researchers retrained a standard SwinV2-based detector on OpenFake and tested it against real social-media content, models trained on older benchmarks failed almost entirely, but the OpenFake-trained detector did not.

OpenFake Arena, a public platform, allows users to try to fool a live detector. Any image that succeeds is folded back into the training data, thus each failure becomes new supervision.

The Dilemma

While the capacity for harm is undeniable, it opens up legitimate creative possibilities for millions.
Content creator Prateek Arora, who uses AI to produce animated narratives, anticipates richer storytelling and novel collaboration models.

Talking about the cutting-edge tools being released today, he told AIM, “I think that’ll really open up the creative ecosystem to many kinds of new ideas, new voices and new talent.” This dramatically reduces production barriers for genres that once demanded prohibitive budgets.

The technology, he argued, allows storytelling to detach from the physical presence of the creator, shifting emphasis toward ideas, structure, and narrative ambition rather than on-camera charisma.

Arora, however, cautioned that people need to be aware of sincere personal responsibility tied to the use of this technology.

“Don’t use it for bad engagement farming. Do it from a place of authenticity. Even if it’s something quick, that’s okay.”

The post Fighting Deepfakes May Not Be a Technology Problem appeared first on Analytics India Magazine.

The Indian deep tech ecosystem has already begun treating 2026 as a delivery deadline. Over the next year, public funding, industrial policy and private capital are expected to converge in ways that could determine whether the country finally moves deep tech out of laboratories and into live operations.

Investors, founders and enterprises agree on one thing: the transition from science to services will not be decided in whitepapers, but on factory floors, warehouses, and supply chains.

But for any ecosystem to roar to life, it needs a spark.

In November, the union government launched the ₹1 lakh crore Research, Development and Innovation (RDI) Scheme Fund to boost private sector R&D in deep tech, AI, and quantum energy. The fund is designed to reduce risk for high-gestation technologies such as semiconductors, AI, robotics, and advanced manufacturing by offering long-term capital rather than short funding cycles, channeled through Anusandhan National Research Foundation.

“The government is actually providing capital for startups. It’s providing capital for venture capitalists,” Sriram Viswanathan, founding managing partner at Celesta Capital, told AIM. “That’s kind of the tailwind you need for an ecosystem to develop.”

Viswanathan argued that patient capital only works if domestic demand exists alongside it. “You cannot expect to only have international or export-related opportunities. You have to become self-reliant,” he said, pointing to procurement and industrial adoption as the missing links.

Celesta in collaboration with other global and Indian venture capitalists, including onboarding NVIDIA, launched the India Deep Tech Alliance committing ₹7500 crores.

What remains uncertain is whether India can bridge the long-standing gap between research and commercial deployment.

From Research to Assembly Line

For early-stage investors like Ankit Kedia, founder and lead investor of Capital A, the bottleneck is not a shortage of innovation but the absence of manufacturing pathways. “You can be a research company for only a certain amount of time,” Kedia said in an exclusive conversation. “In the VC world, the scale will come from manufacturing.”

Capital A’s Fund II, with a $50-million corpus to fund manufacturing, deep tech, climate and fintech startups, focuses on taking products “from the lab to the shop floor”, with a longer view on timelines. “You have to give it a minimum of three years to come to the shop floor and then three to six years to scale,” he said. Shorter cycles, he added, do injustice to the manufacturing industry.

Sunil Shekhawat, founder and CEO of deep tech enablement platform and accelerator SanchiConnect, agrees. “If something can be developed in 18 months, that needs to be questioned.” In his view, the real test is the ability to replicate and the time to commercial maturity.

This thinking aligns with the government’s own timelines. India’s semiconductor push, under the India Semiconductor Mission, has repeatedly referenced 2026 as a north star for establishing meaningful domestic capacity across design, packaging, and manufacturing.

While large fabs remain capital-intensive, policymakers have signalled that design-led and fabless startups are expected to mature alongside infrastructure build-out.

Yet the middle layer remains fragile. Many MSMEs operate between proof-of-concept and industrial scale, unable to attract venture funding or bank credit. “They can’t scale up exponentially,” Kedia said. “That layer in between is protecting the entire middle economy of India.”

Beyond the Metros

Bridging that gap increasingly depends on ecosystem infrastructure rather than individual capital cheques. SanchiConnect’s Shekhawat said the challenge is not matchmaking but proximity.

SanchiConnect brings together founders, corporate mentors, investors, and incubators, many of whom are outside major metros. Over three years, the platform has mobilised more than ₹250 crore into early-stage deep tech companies. “70% of the people are from Tier 2 cities,” Shekhawat said.

That geographic spread matters, as manufacturing capacity does not cluster as software talent does. It also aligns with state-level skilling and industrial programmes expected to mature through 2026, particularly in electronics, drones, and advanced manufacturing.

Enterprises Push Adoption

While investors debate gestation, enterprises are already deploying deep tech in frontline operations. Zebra Technologies, which works across logistics, manufacturing and retail, sees India moving from pilots to scale.

“Over the past two years, we’ve already seen customers transition from piloting emerging technologies to deploying them at scale across factories, warehouses and retail networks,” said Subramaniam Thiruppathi, director for India and the subcontinent at Zebra Technologies.

By 2026, Zebra expects AI to be embedded into everyday decision-making for frontline workers.

According to an IDC-UiPath study, 40% of Indian organisations have implemented agentic AI and another 50% are planning to do so by 2026. Much of this shift relies on on-device intelligence rather than cloud dependence—a response to inconsistent infrastructure.

This enterprise momentum mirrors the government’s policy calendar. The India-AI Impact Summit, scheduled for February 2026, is positioned as a national forum to align policy, enterprise adoption and startup deployment around AI.

Pre-summit accelerator programmes are already underway, signalling that 2026 is being treated as an execution year rather than just another year for deliberation.

Unresolved Constraints

Despite policy momentum, structural gaps persist. Hardware startups still rely heavily on imported components, a challenge shaped by cost, certification, and access. The government’s extension of the Import Management System for IT hardware until December 2026 reflects this tension between domestic manufacturing ambition and near-term dependence.

Kedia traces the problem back to exposure. “If we are able to expose the specifications, the high-quality expectations, I’m sure MSMEs will make the effort,” he said. Without that visibility, localisation remains slow.

Viswanathan framed the issue more bluntly. “Most successful companies are ‘overnight successes’ in ten years,” he said. “What happens in the middle pretty much defines it.”

That middle is where 2026 will be tested. If RDI capital reaches companies on time, enterprises continue pulling deep tech into operations, and manufacturing capacity grows beyond policy announcements, India may finally narrow the gap between ambition and execution.

If not, 2026 will pass as another milestone on paper. The difference this time is that the shop floor is no longer waiting quietly.

The post Will India Finally Hit the Deep Tech Sweet Spot in 2026? appeared first on Analytics India Magazine.

Meta has rolled out a new software update for its AI-powered smart glasses, improving audio quality through noise filtration and enhancing the audio streaming experience with Spotify using multimodal AI.

The v21 update introduces “Conversation Focus,” a feature that amplifies a speaker’s voice in noisy environments. First announced at Meta Connect earlier this year, the feature uses open-ear speakers built into Ray-Ban Meta and Oakley Meta HSTN glasses to enhance speech clarity. The system selectively amplifies the voice of the person a user is speaking with, helping distinguish conversation from background noise in places such as busy restaurants, trains, or crowded events. Users can adjust amplification levels directly from the glasses or through device settings, depending on their surroundings.

The update also introduced Meta’s first multimodal AI music experience in partnership with Spotify. By combining on-device vision with Spotify’s personalisation engine, users can ask Meta AI to play music that matches what they are looking at, blending visual context with individual listening preferences to create moment-specific soundtracks.

In a move that strengthens its India-focused AI strategy, Meta has added Telugu and Kannada language support to Ray-Ban Meta and Oakley Meta HSTN glasses. The rollout enables fully hands-free interaction with Meta AI in two additional regional languages, making the devices more accessible and natural to use for millions of users across the country.

With this addition, Meta AI’s multilingual footprint in India now extends beyond English and Hindi, reflecting a broader push to localise AI-powered wearables for diverse linguistic communities.

The new features are rolling out gradually, starting with users enrolled in Meta’s Early Access Programme, with wider availability expected over time.

The post Meta Ushers Spotify Integration, Kannada & Telugu Support, Noise Filtration to AI Glasses appeared first on Analytics India Magazine.

IBM has committed to skill 5 million learners across India in AI, cybersecurity and quantum computing by 2030, as part of its workforce development efforts in the country.

The initiative is part of IBM’s SkillsBuild tech learning ecosystem and will focus on students and adult learners across schools, universities, and vocational ecosystems. The company said the India programme aligns with its focus on employability, responsible AI education and early exposure to computational thinking, as demand for advanced digital skills continues to grow across sectors.

The initiative, announced in Delhi, aims to expand access to education in emerging technologies by working with academic institutions and regulators, including the All India Council for Technical Education. The company also plans to support hands-on learning through curriculum integration, faculty enablement, hackathons, and internship pathways.

It will also continue work at the school level by co-developing AI curricula for senior secondary students and providing teaching resources, including project guides and explainer modules.

“India possesses the talent and ambition to lead the world in AI and quantum,” said Arvind Krishna, chairman, president and chief executive officer of IBM, in a statement. “Our commitment to skill five million people is an investment in that future,” he said, adding that broader access to advanced skills would help students build and innovate.

In a previous interaction, L Venkata Subramaniam, who served as the quantum India leader at IBM, told AIM that IBM is co-developing 11 textbooks on quantum computing with IITs, startups, and other partners, with over 100 colleges already signed up. This is part of a nationwide rollout of an undergraduate minor programme in quantum technologies.

At the centre of the effort is IBM SkillsBuild, a digital learning platform that offers more than 1,000 courses across technology and workplace skills. IBM said the platform has reached over 16 million learners globally and supports its broader goal of training 30 million people worldwide by 2030, with India playing a key role.

The post IBM to Skill 5 Mn Indian Youth in AI & Quantum by 2030 appeared first on Analytics India Magazine.

Databricks is operating at a scale most companies reach only after going public. With a $4.8 billion annual revenue run rate, positive cash flow, and a valuation of $134 billion, the data platform is challenging long-held assumptions about when—or whether—high-growth tech companies need an IPO.

The data platform recently announced raising more than $4 billion in a Series L round, a scale of late-stage funding that remains rare, as demand for AI and data platforms continues to grow.

Commenting on the company’s prolonged stay in the private markets, Vasuman M, founder of AI engineering firm Varick Agents, joked on X, “Imagine joining Databricks in 2017 at its Series D, thinking the IPO was just around the corner, and then nine years later the company is raising a Series L.”

Only SpaceX, OpenAI, Anthropic, and xAI are higher-valued private US companies, and three of them are looking to go public.

Databricks’ funding comes after the company reported more than 55% year-over-year growth in the revenue run-rate in Q3, projecting over $1 billion in revenue from AI products and a similar amount from its data warehousing business.

The company said it has remained cash-flow positive over the past 12 months, a rarity among high-growth tech firms in the AI space, investing heavily in data centres.

Interestingly, since its founding, the company has raised about $19 billion in private funding to date, according to Tracxn, far exceeding its rival Snowflake’s $1.4 billion in capital raised before its 2020 IPO. That puts Databricks’ private fundraising at roughly 13 times higher than Snowflake’s, even as the two companies have a similar revenue run rate.

Snowflake is currently valued at about $76 billion, significantly below Databricks.

Why Not IPO?

In a recent interview with CNBC, Databricks CEO Ali Ghodsi said that remaining private has given the data company the flexibility to keep investing through market cycles without the pressure of short-term expectations.

“If we had gone public earlier…I don’t think we would have been at this growth rate right now,” he said, adding that staying private allowed it to invest in areas such as AI agents and its database business, including Lakebase.

Over the past few years, Databricks has made several large acquisitions—spending about $1.3 billion on GenAI startup MosaicML, more than $1 billion on data management platform Tabular, and roughly $1 billion on Neon—as it builds out its AI and data platform capabilities.

Ghodsi said Databricks has been free-cash-flow positive but is intentionally reinvesting rather than optimising for margins. “We’re an efficient business, but we’re reinvesting it. We just want to stay at break-even and invest it back in agents and databases,” Ghodsi said. He further added that the company is betting on what it calls a new category of “data intelligent applications.”

On a potential IPO, Ghodsi said going public remains an option, possibly as early as next year, but cautioned against listing in a market that could limit long-term innovation. “I don’t want to take the company public and then have the market demand 30% EBITDA,” he said.

Addressing concerns about an AI bubble, Ghodsi said AI adoption is structural rather than cyclical. “People are not going to stop using AI,” he assertedsaid, pointing to growing automation in software development and enterprise workflows that will continue to drive demand for data platforms built to work with AI agents.

What Makes Databricks Special?

Databricks’ co-founder and chief architect, Reynold Xin, in a post on X, said that the new funding challenges the conventional belief that “it would take 5+ years to build a new database, just to release one.”

He traced the origins of the data warehousing business to four years ago, when Databricks’ DBSQL, then still in preview, topped the official TPC-DS 100TB benchmark, outperforming a previous best by 2.2x, including a 12x advantage over Snowflake at the time.

Databricks, Xin added, continues to hold the top spot on the benchmark.

According to Xin, Databricks built the business by assembling a dedicated engineering team and introducing the Lakehouse architecture, which the company said combines the openness and flexibility of data lakes with the performance of traditional data warehouses. He said the Lakehouse model has since become “the standard for data infrastructure,” with enterprises increasingly migrating away from legacy warehouses.

Xin also said Databricks is expanding beyond analytics into operational workloads, arguing that online transaction processing (OLTP) systems, managing day-to-day business operations, are ready for a similar disruption. He said a significant portion of the founding team is now working on Lakebase, a new OLTP database category that separates storage in the data lake from compute.

As the year draws to a close, the company announced a set of product updates to Lakebase, including autoscaling that adjusts compute based on workload, scale-to-zero support with automatic resume in milliseconds, and instant provisioning that allows new database instances to be created in seconds.

The release also introduces instant database branching for git-like development, testing, and staging workflows, along with automated backups and point-in-time recovery.

Additional updates include support for open-source relational database Postgres 17 alongside Postgres 16, expanded storage capacity up to 8TB for production workloads, and a new Lakebase user interface designed to simplify common operational tasks.

The next phase of growth will test whether Databricks can convert ambitious product bets into broad enterprise data adoption.

The post After $4 Bn Databricks Haul, is IPO Still Endgame for AI Companies? appeared first on Analytics India Magazine.

ByteDance has signed a binding agreement to transfer control of TikTok’s US operations to a consortium of predominantly American investors led by Oracle and private equity firm Silver Lake, Axios reported, citing an internal company memo. This brings an end to years of regulatory uncertainty around the short-video app.

The deal also includes Abu Dhabi-based investment firm MGX and will result in the creation of a new US-based entity that will oversee TikTok’s business in the country.

The transaction is expected to close on January 22, TikTok CEO Shou Zi Chew said in the memo. The new entity, to be called TikTok USDS Joint Venture LLC, will be owned 45% by Oracle, Silver Lake and Abu Dhabi-based MGX, while affiliates of existing ByteDance investors will hold nearly one-third. ByteDance will retain close to a 20% stake.

The joint venture will assume responsibility for US data protection, algorithm security, content moderation and software assurance. As part of the agreement, the company will retrain TikTok’s content recommendation algorithm using US user data to ensure the feed is insulated from external influence, the memo said.

Oracle will serve as the trusted security partner, responsible for auditing and validating compliance with national security terms once the deal closes.

Following completion, the US joint venture will operate independently with authority over sensitive US operations, while TikTok’s global entities will continue to manage product interoperability and certain commercial activities such as advertising, marketing and e-commerce.

The deal values TikTok’s US business at around $14 billion, according to a source cited by Axios.

The agreement follows a deal in principle reached in September between the White House and the Chinese government to sell TikTok’s US operations.

Pressure on ByteDance to divest began in 2020, when then-president Donald Trump issued an executive order seeking a sale on national security grounds. Congress later passed legislation in 2024 to ban the app unless it was sold, a law upheld by the Supreme Court earlier this year.

The post Oracle and Silver Lake Take Control of TikTok’s US Business appeared first on Analytics India Magazine.