Reuters traced $2.3bn in Iranian exchange Nobitex’s flows since 2023 to Tron and BNB Chain, the blockchains established by World Liberty Financial’s two most prominent early backers. No party at WLF has been accused of knowing about it.

A Reuters investigation published on Monday documents that Nobitex, Iran’s largest crypto exchange, has processed at least $2.3bn since 2023 on the Tron and BNB Chain blockchains. Tron was founded by Justin Sun.

BNB Chain was developed by Binance, the exchange owned by Changpeng Zhao. Both Sun and Zhao are the two most prominent early backers of World Liberty Financial, the crypto firm co-founded by Donald Trump and his family. There is no suggestion that the Trump family knew about Nobitex’s use of either network.

Reuters’ breakdown of public blockchain data, sourced from Arkham, puts about $2bn of the Nobitex flow on Tron and $317m on BNB Chain since 1 January 2023, with $22.6m on BNB Chain and $550,000 via Tron since the Iran war began in February.

Four crypto analysts called the calculation sound; independent investigator Rich Sanders said the true figure was probably higher, since public flows are visible only for known Nobitex addresses and the exchange has admitted to switching addresses to avoid tracing.

The principals’ positions on the news are all on the record. The White House response, from spokeswoman Anna Kelly, described the article as ‘bizarre attempts to link President Trump to Iran’s banking system’ that were ‘totally laughable’.

A spokeswoman for World Liberty said the company ‘has no relationship with Nobitex and follows U.S. law’, adding that ‘World Liberty does not own, operate, or control Tron in any way, and has no authority over transactions conducted on it’. Nobitex said any illicit funds passed through the exchange ‘without management approval or awareness’.

A Tron spokeswoman said the network ‘is a technology provider’ that cannot ‘monitor and investigate every user and every transaction’. Binance said it was ‘an initial contributor and incubator’ of BNB Chain rather than its operator.

Binance’s corporate-structure caveat sits inside a longer record. Abu Dhabi filings reviewed by Reuters show Zhao as the sole listed shareholder of BNB Chain Technology Holding Limited, the entity to which BNB Chain’s operations were transferred in 2023.

Reuters’ 2022 reporting found about $7.8bn in crypto flowed between Nobitex and Binance from 2018 to 2022, roughly three-quarters in Tron’s native asset, with Nobitex actively encouraging clients to use Tron to trade ‘without endangering assets due to sanctions’.

The Trump family’s commercial position in WLF is the structural fact underneath the reporting. Sun’s portfolio of 4 billion WLFI tokens is worth roughly $266m, according to Reuters’ calculations; Binance now holds $3.8bn of the Trump-family token.

Abu Dhabi’s MGX bought a $2bn stake in Binance in early 2025 and announced the transaction would be settled in WLF’s USD1 stablecoin. Trump’s October 2025 pardon of Zhao, wiping his federal conviction for failing to maintain an effective anti-money-laundering programme, sits alongside that commercial timeline.

Binance and Zhao’s lawyers have said there was no connection between the USD1 deal and the pardon.

The narrower money-laundering arc is now well-documented. Tether froze multiple Nobitex wallet addresses at the request of Israel’s National Bureau for Counter Terror Financing.

Elliptic and the same two Iran specialists reported in January that the Central Bank of Iran, sanctioned by the US in 2019 over alleged IRGC and Hezbollah financing, bought more than $500m of tether via Tron between November 2024 and June 2025, of which roughly $347m was routed to Nobitex in the first half of 2025.

The central bank also converted holdings into other coins and moved them across BNB Chain to obscure the trail.

What the United States has not done is sanction Nobitex itself. Reuters notes plainly that it could not determine the reason. Watchdog group Public Citizen, in a parallel report, has called the WLF/Binance/Iran chain a ‘conflict coin’ and pressed for Treasury and DOJ investigations.

Senators Elizabeth Warren and Jack Reed have separately sought formal probes into WLF’s sanctions controls, which the company has said are ‘the highest standard in the industry’.

The named principals had already fallen out before the Reuters investigation. Sun sued World Liberty in April over allegedly frozen assets; WLF countersued in early May, alleging defamation.

The blockchain data does not, on the investigation’s own framing, prove coordinated conduct. It documents that two of the people most exposed to the financial upside of the Trump family’s crypto company are also the architects of the two networks that Iran’s largest sanctioned-flow conduit runs on.

Ubotica will deploy and operate AI models directly in orbit using Novi’s smart-satellite constellation.

Dublin-based NASA and the European Space Agency collaborator, Ubotica, is partnering with Texas’ Novi Space to deliver real-time intelligence from the Earth’s orbit.

Novi provides computing technology for spacecrafts, alongside a constellation of multi-sensor edge-processing satellites linked to an intelligence management platform.

The open-access platform allows companies to build AI-powered applications in space, including improving Earth observation, geospatial intelligence and autonomy.

The collaboration integrates Ubotica’s AI platform with Novi’s ‘Genie’ smart-satellite constellation and platform to enable Earth observation data to be processed directly onboard satellites, unlike traditional systems that transfer data to Earth before analysis.

Through the partnership, Ubotica will deploy and operate AI models directly in orbit through the Genie multi-sensor satellite platform.

The Dublin start-up’s ‘Space:AI’ platform generates analysis within 90 seconds from when it begins processing, it said, resulting in lower latency and bandwidth costs translating to savings.

According to the company, in a single observation of a Singapore port, the platform processed hundreds of vessels and detected those operating dark in under two minutes. The company has deployed its AI capabilities on numerous missions, including its own CogniSAT-6 satellite.

“Our collaboration with Novi brings more AI-enabled Earth observation capacity into orbit. By combining Space:AI with Genie’s onboard compute, we’re shifting satellites from data collectors to intelligent agents, delivering insights in minutes rather than days,” said Ubotica co-founder and chief technology officer Dr Aubrey Dunne.

“That capability underpins our live maritime intelligence service and unlocks new operational models for time-critical surveillance.” The start-up won the SpaceNews Icon Award for Space AI Partnership last December, alongside NASA JPL and Open Cosmos.

Michael Bartholomeusz, the CEO of Novi Space added: “Partnering with Ubotica allows us to demonstrate the full potential of Genie – delivering real-time intelligence from orbit, not just data. This is a fundamental shift in how space-based systems create value.”

Ubotica and NASA announced a partnership last month to onboard the start-up’s AI platforms in a test mission to demonstrate autonomous intelligent satellite networks. In February, the company was among the first chosen for involvement in Ireland’s European Space Agency Phi-Lab at Irish Manufacturing Research.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

AMD has launched the Ryzen PRO 9000 Series processors for commercial desktops and workstations. With these processors, the company offers enterprise-level 3D V-Cache capabilities to enhance performance for demanding workloads. AMD designed these processors for professionals working in content creation, architecture, engineering, and design.

These processors are the latest in AMD’s lineup, designed for commercial desktops and professional workstations. Along with performance upgrades for heavy workloads, the lineup also includes enterprise security and long-term platform support features.

3D V-Cache for Professional Workloads

3D V-Cache will be introduced in enterprise workstation processors with the upcoming Ryzen PRO 9000 Series. With increased cache memory, the company aims to improve processor performance when handling intensive tasks or applications. The increased cache memory will enhance the performance of applications working on large files by making it easier for them to access them. The company states that the technology aims to increase processor efficiency in professional environments.

The Ryzen PRO 9000 Series is designed for professionals such as creators, architects, engineers, and designers who rely on heavy-duty professional software. According to AMD, the chips will provide improved performance during the editing and encoding of videos in 4K and 8K resolutions, along with compositing performance in media workflows. The processors are also intended for architects and construction professionals, including BIM and 3D modeling. Manufacturing professionals working with CAD models and simulations will benefit from the performance of this series of chips.

Performance and Security

The AMD Ryzen PRO 9000 Series comes in various configurations designed for desktop computers and business workstations. These CPUs are offered in six-core, eight-core, 12-core, and 16-core varieties with several thermal design power (TDP) choices. AMD has created high-performance CPUs for individuals who require reliable performance on challenging projects.

Along with performance upgrades, the Ryzen PRO 9000 Series also focuses on enterprise reliability and security. As part of the AMD PRO platform, the processors include advanced security protections and manageability tools for IT departments. AMD says the platform supports long-term business deployments with stable and consistent performance.

Availability Details

According to AMD, the Ryzen PRO 9000 Series processors will arrive in the second half of 2026. Lenovo has already confirmed that it will feature the new processors in its ThinkStation P4 workstation at NXTBLD. AMD may also announce additional OEM partners and systems closer to launch.

AMD Ryzen PRO 9000 Series Specifications

Databases

Cloud giant says humans remain accountable, even when code gets an assist from the machines

Google is encouraging its database developers to lean “heavily” on AI coding tools as it ramps up contributions to open source projects such as PostgreSQL.

Earlier this year, Google announced a raft of new contributions to PostgreSQL, the open source database that has become a popular RDBMS for developers building new applications in the cloud.

Sailesh Krishnamurthy, VP of Databases, Google Cloud, told The Register that the company was using AI coding tools to accelerate its contributions to open source database systems, although each developer remains responsible for their individual contributions.

“We do encourage folks to use AI heavily ,” he said. “We are seeing huge amounts of productivity improvements internally. In the end, we have individual engineers take accountability for our contributions. Whether you have a piece of code that is completely drafted by AI, or not even part of what you’re pasting into your development environment, you have a whole spectrum where AI is used in different places. Either way, the accountability remains on behalf of the person who’s done it.”

AI coding tools can be especially suited to developing contributions to open source projects because the codebase is publicly available and has been used to train the generative models, he said.

“That’s how models have a better sense of the code, as opposed to many proprietary pieces of code, which are inside the firewall.”

PostgreSQL was designed to be extensible. As such, it can be a system well suited to vibe coding to get new ideas off the ground quickly, Krishnamurthy said.

“The sweet spot is where you have maybe an interesting academic idea that is well understood, and you have a codebase that’s well understood, and you’re trying to say, well, I want to take this idea and I want to take this piece of code and build an extension for it. That’s a great example where you have something isolated – the blast radius is small – and you can go and use AI to interpret the code. Our own engineers are using AI quite heavily, but also judiciously.”

PostgreSQL became the most popular database among developers in 2023, according to the Stack Overflow survey. The trend owes a great deal to the plethora of PostgreSQL database services out there, not least from the big three cloud providers, which have ramped up investment in the open source system.

Last year, Microsoft contributed pg_documentdb_core, a custom PostgreSQL extension that enables support for Binary JavaScript Object Notation (BSON, a binary-encoded serialization of JSON documents), and pg_documentdb_api, a data layer providing MongoDB-compatible commands for create, read, update and delete (CRUD) operations, queries, and index management. The extensions are set to run on the Azure Cosmos DB PostgreSQL database service and offer a document-store-style database to rival MongoDB.

Microsoft has also announced a distributed PostgreSQL database service called HorizonDB.

Krishnamurthy said: “The industry at large is investing heavily in PostgreSQL. We see this across the board, whether it’s customers, whether it’s digital native services, and certainly we see the migrations coming from commercial databases. It is also a broad industry trend of PostgreSQL as a layer, no matter where data is being stored.”

As such, Google has contributed new code to the project, with the engineering effort focused on advancing logical replication. Contributions included Automatic Conflict Detection, designed to allow the replication worker to automatically detect when an incoming change (Insert, Update, or Delete) conflicts with the local state; and logical replication of sequences.

Demand for PostgreSQL services is coming from migrations as well as new applications, Krishnamurthy said. Customers are ditching Oracle, Microsoft SQL Server, and IBM Db2, as well as other legacy systems, including Sybase and Informix.

Research from Gartner earlier this year shows that of the leading database vendors 15 years ago – Oracle, IBM, Microsoft, and SAP – only Microsoft has grown its market share since. As well as its own database systems, Microsoft offers PostgreSQL and MySQL services, as does AWS, the leading database vendor. Oracle remains third, ahead of Google, and that position seems unlikely to change soon. Nonetheless, with all the major cloud vendors contributing to open source database projects such as PostgreSQL, momentum is slowly shifting. ®

Google has always been ahead of the curve when it comes to protecting Pixel users from spam calls, and it looks like the company isn’t done yet. According to a recent teardown of the Google Phone app by Android Authority, Google is working on a new phone number spoofing detection feature.

What is phone number spoofing?

Phone number spoofing, also known as caller ID spoofing, is when a scammer tricks your phone into displaying a familiar or saved contact’s number, even though the call is actually coming from a completely different number. 

As users are more likely to pick up a call if it looks like it’s coming from family members, friends, or authorized personnel, like a doctor or a bank representative, phone number spoofing is on the rise in the scam chart. It has become a surprisingly common tactic and one that has caught a lot of people off guard.

So what is Google doing about it?

Android Authority cracked open version 222.0.913376317 of the Google Phone app and found strings of code that point to an upcoming spoofing detection system. One of the strings reads, “Someone may be pretending to call from your contact’s number,” and another suggests that users will have the option to hang up the call immediately.

It’s not entirely clear how Google plans to detect spoofed numbers, but the timing is interesting. Only a few days back, Google announced a slew of security features, including verified financial calls, OTP protection, real-time malware detection, APK scanning in Chrome, and more.

With the new call spoof detection feature and existing spam call protections, including Call Screening and spam detection, the Pixel phones have become the best anti-scam smartphones. There’s no word yet on when this feature will roll out, but it’s good to know Google is working on it.

The Tampa-based, Sun Capital-owned software services firm is folding the SAP- and IWG-trusted UK digital agency into its Digital Experiences practice. Deal terms were not disclosed.

Exadel, the Tampa-based software-development and consulting firm, has acquired Tangent, the London-based digital experience consultancy, the company said on Monday. Terms were not disclosed.

Tangent will continue to operate under its existing brand inside Exadel’s Digital Experiences practice, with chief executive Leigh Gammons moving into a managing director and senior vice president role to lead the business.

On Exadel’s own framing, the deal is about pairing two halves of an enterprise transformation engagement that have traditionally lived in different vendor categories.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Exadel sells what it calls AI-native engineering: data infrastructure, applications, and the back-end work of running an enterprise’s technology stack, on a 2,000-plus headcount across the US, Europe and LATAM.

Tangent sells the front-end discipline, including UX, product, web experience and MarTech engineering, on a smaller boutique footprint built up around what its website describes as enterprise digital-product work. The acquisition pulls strategy, design and engineering inside one contract.

‘Brands increasingly win or lose based on the AI-driven digital experiences they provide to customers,’ said James Dalziel, Exadel’s chief operating officer, in the statement.

‘By bringing Tangent into the organisation, we are fortifying our ability to help global clients not only design exceptional experiences, but also continuously optimise and scale them through AI.’

Gammons described the value of the combination from the other direction: ‘Companies are demanding more than great digital experiences. They need to provide experiences that can constantly evolve and drive measurable outcomes.’

Tangent has been operating since 2001 and counts SAP, IWG, and UK Power Networks among its enterprise clients, according to its own published materials, with the Exadel release also naming New Balance and Vodafone.

Its team is London-headquartered with a Newcastle office and a delivery footprint across Spain, South Africa, Poland, Egypt, and Pakistan. Gammons joined the agency as chief executive after leaving a senior role at WPP.

Exadel’s M&A appetite has a recognisable shape. The company is owned by an affiliate of Sun Capital Partners after a take-private transaction, and has spent the past few years adding capability through bolt-ons.

Its prior acquisitions include Motion Software, CPQi and Coppei. Tangent is the latest in that arc, and the first explicitly aimed at the design-and-strategy front of the enterprise stack rather than the engineering or sector-specific back.

The two companies have also announced a joint AI accelerator programme for enterprise clients, framed as a way to take engagements from ‘AI ambition to real-world delivery’.

The structural detail, pricing, and pilot customers have not been published. The framing positions Exadel as an AI-native alternative to the Big Four and the global systems integrators on engagements where Tangent’s experience-design front-end has historically been outsourced to a different vendor.

The acquisition lands inside a broader recalibration of the enterprise-services category that has been visible for several quarters. AI-agent products from the foundation labs have started to reach directly into the workflows that consultancies have traditionally billed for.

Anthropic shipped ten financial-services agent templates earlier this month, pulled Moody’s data inside the workspace, and built distribution through Microsoft 365 and Snowflake.

SAP unveiled an Autonomous Enterprise framework with more than 200 AI agents at Sapphire on a co-development with Anthropic. The competitive question for a services firm with Exadel’s profile is no longer whether the AI side of the stack will be the most valuable; it is whether the integrator that can plug the model layer into the customer experience layer end-to-end retains pricing power against the model layer itself.

Whether the Exadel-Tangent combination has the scale to be that integrator is the question the next 18 months of customer wins will settle.

The ‘Exadel Colleague’ AI delivery product, which Exadel launched last month, is the company’s bet that its engineering side will not be commoditised by the models.

Tangent’s customer roster is what determines whether the design side, attached to that engineering side, gives the combined business a contract-by-contract structural advantage in front-end-heavy enterprise work.

Google‘s annual Google I/O developer conference is almost upon us, and as well as getting new features for Android phones and a better look at those new Googlebooks, as a wearables enthusiast I’m curious to see what happens with Wear OS 7.

Because Google I/O is primarily for developers, we should get a better look at the latest slate of operating systems and AI powers that devs can use to design new apps and features for the likes of the Google Pixel Watch 4, Samsung Galaxy Watch8, and more upcoming, unannounced devices.

Commencement season has come around again — and this year, a couple speakers have discovered that it’s tough to get graduating students excited about a future shaped by artificial intelligence.

Last week, Gloria Caulfield, an executive at real estate firm Tavistock Development Company, gave a speech at the University of Central Florida acknowledging that we’re living in a time of “profound change,” which can be both “exciting” and “daunting.”

“The rise of artificial intelligence is the next industrial revolution,” Caulfield declared — prompting the students in the audience to begin booing, getting louder and louder until Caulfield chuckled, turned to the other speakers, and asked, “What happened?”

“Okay, I struck a chord,” she said. Caulfield then tried to resume her speech, saying, “Only a few years ago, AI was not a factor in our lives” — only to be interrupted again by the audience, this time by their loud cheers and applause.

Former Google CEO Eric Schmidt faced a similar response when he brought up AI at a University of Arizona speech on Friday.

In Schmidt’s case, the criticism actually began before the speech itself, with some student groups calling for him to be removed as commencement speaker due to a lawsuit in which a former girlfriend and business partner accused Schmidt of sexual assault. (He has denied the allegations.) According to a local news report, the booing began even before Schmidt took the stage.

But Schmidt also got loud boos when he told students, “You will help shape artificial intelligence.” The booing was persistent enough that Schmidt tried to speak over it, insisting, “You can now assemble a team of AI agents to help you with the parts that you could never accomplish on your own. When someone offers you a seat on the rocket ship, you do not ask which seat, you just get on.”

To be fair, AI isn’t becoming a third rail at every graduation ceremony. Nvidia CEO Jensen Huang recently spoke at Carnegie Mellon’s commencement, and he didn’t seem to get any audible pushback when he said that AI has “reinvented computing.”

Still, it’s not exactly surprising to find some students in a booing mood. In a recent Gallup poll, only 43% of Americans aged 15 to 34 said it’s a good time to find a job locally, a steep drop from 75% in 2022. 

That pessimism isn’t solely a response to the rise of AI (a shift that even some software engineers are worried about), but journalist and tech industry critic Brian Merchant suggested that for many students, AI has become “the cruel new face of hyper-scaling capitalism.”

“I too would loudly boo at the prospect of this next industrial revolution if I was in my early twenties, unemployed, and had aspirations for my future greater than entering prompts into an LLM,” Merchant wrote.

Even when graduation speeches didn’t mention AI explicitly, “resilience” was a recurring theme this year. Schmidt himself acknowledged that there is “a fear in your generation that the future has already been written, that the machines are coming, that the jobs are evaporating, that the climate is breaking, that politics are fractured, and that you are inheriting a mess that you did not create.”

Caulfield, meanwhile, might also have misread her audience of arts and humanities graduates. One student said that before mentioning AI, Caulfield already started to lose them with her “generic” praise of corporate executives like Jeff Bezos.

Another graduate, Alexander Rose Tyson, told The New York Times, “It wasn’t one person that really started the booing. It was just sort of like a collective, ‘This sucks.’”

Best Buy found a way to stay alive and even thrive as contemporaries like Circuit City fumbled in the age of Amazon. Now it’s one of the biggest electronics retailers on the web, leveraging a quality online shopping experience with brick-and-mortar outfits so you can choose to wait for delivery or head to your nearest store and get your new phone, TV, laptop, or audio gear in a hurry. With a Best Buy discount code or Best Buy promo code, you can save on the company’s already competitive pricing. Here are some hot deals currently available.

Get 10% Back in Rewards With the My Best Buy Credit Card

This Best Buy Credit Card deal is a solid offer for anyone shopping big on electronics like OLED TVs, headphones, soundbars, and other gear. For everything you buy on your first day of purchases, you’ll get 10% back, including 2.5 bonus points and an additional 5% back in rewards. There are, of course, stipulations and terms. You’ll need to make your purchases within 14 days of opening a new account, you can’t combine these deals with other offers, and points aren’t awarded on promotional credit purchases. The deal is available in select stores, and online offers “may vary.”

You can also choose financing options like 12-month financing on storewide purchases totalling $299 and up, 18-month financing on appliances and grills totaling $599 and up, or 24-month financing on unlocked phone purchases totaling $649 and up.

Save up to 60% With Top Deals at Best Buy

Best Buy offers some sweet daily deals, including Top Deals that could save you up to 60% off retail pricing, and 24-hour discounts on a rotating deal of the day to help you save on gear like laptops, TVs, appliances, and other products. These deals switch swiftly, so be sure to check back often so you can jump on the next big discount.

Join My Best Buy for Free Shipping, Member Prices, and More

If you’re in it for the long haul, it’s worth signing up for My Best Buy. It’s free and you get all sorts of Best Buy discounts and promos, along with multiple other benefits, starting with Free Standard Shipping across all levels.

Move up to the My Best Buy Plus ($50/year) or My Best Buy Total ($180/year) will get you benefits like two-day shipping, exclusive member prices and deals on thousands of items, access to exclusive sales events and products, and a 60-day return window on the majority of products.

The My Best Buy Total plan offers a few other extras, including protection plans like AppleCare+, free in-store and online support services for computers and tablets, 24/7 tech support, 20% off repairs, and more.

Score Bonus Perks as a My Best Buy Plus or Total Member

Those aren’t the only perks of jumping on one of Best Buy’s paid membership plans. Join up with the My Best Buy Plus or My Best Buy Total plans and get a slew of other Membership Deals, including a daily chance to win $1,000 in My Best Buy certificates (through May 2).

The paid membership deals also unlock all sorts of subscription deals, including , 30 days of free Fubo Pro, a 2-month trial of Tastemade+ recipes, 6 months of free SiriusXM All Access satellite radio, 3 months of free YouTube Premium, and 60 days of free LifeLock ID theft protection (along with 75% off on your first year’s subscription). You can also get a 90-day free trial for McAfee Privacy and Identity Guardian Online protection (or 80% off a 1-year subscription), and a month of Discord Nitro for free, to enhance your Discord experience with personalized profiles, animated emojis, and more.

Use the Best Buy Price Match Guarantee to Always Pay the Lowest Price

Finally, there’s no better way to stretch your dollars than to leverage hot deals on items like TVs, monitors, computers, and other electronics from major competitors to save with the Best Buy Price Match Guarantee. The deal works like this: identify an identical product from a Qualified Competitor (which, sadly, excludes Amazon), then connect with Best Buy over Best Buy Chat online or by calling 1-888-BEST BUY to save.

AI companies are not just changing the way people work. They are changing the kinds of roles that exist. Org charts are morphing as an entirely new class of jobs emerges, some with titles that did not exist two years ago, others that represent old professions reborn inside the technology industry. The hiring sprees stand in stark contrast to the layoffs that many of the same companies are citing AI as the justification for.

The range is striking. Companies are looking for everything from one of the oldest intellectual pursuits, philosophy, to an entirely new category of work spawned by generative AI coding tools, the professional vibe coder. Between those poles sit forward deployed engineers, AI accelerators, evangelists, gig workers training models, and a growing class of C-suite officers whose entire job is to make sure the rest of the company uses AI.

The forward deployed engineer is the hottest role in the category. Popularised by Palantir in the 2010s, the job embeds a specialised engineer directly with a customer to deliver tailored AI solutions rather than off-the-shelf software. Indeed data shows that job postings for forward deployed engineers in January 2026 were roughly 19 times the volume of the year before. Palantir CEO Alex Karp has compared the role to a seasoned waiter in a French restaurant, combining deep product knowledge with exquisite service. OpenAI, Anthropic, Google, Amazon Web Services, and Palantir are all hiring, with starting salaries ranging from $115,000 to more than $200,000. Salesforce’s projected $300 million in Anthropic token spending this year illustrates the scale of enterprise AI adoption these engineers are being hired to support.

The AI evangelist is a different kind of hire. Anthropic is looking for a “Claude Evangelist,” someone who will serve as the company’s face in the startup ecosystem, combining at least seven years of founder-builder experience with developer-facing credibility. The role pays $240,000, significantly more than the $106,000 average for a US director of communications, according to Indeed. OpenAI has tripled the size of its communications team. Adobe is hunting for a “Business Architect & AI Evangelist.” The underlying logic is that AI products are too complex and too consequential to sell through conventional marketing. They require people who can explain, demonstrate, and build trust in person.

The AI philosopher may be the most unexpected entry. Anthropic has a resident philosopher. So does Google DeepMind. Both positions focus on ensuring AI models are aligned with human values. Anthropic publishes a Constitution for Claude, a detailed description of the values it wants its AI to have, and the philosophical work behind it is not decorative. Google DeepMind recently sought an emerging impacts manager in AI ethics and safety with a base salary of $212,000 to $231,000. Philosophy departments that have spent years defending their enrolment numbers now have a direct pipeline into technology companies paying more than twice the median salary for the discipline.

The internal AI accelerator is the role that most directly confronts the tension between AI hiring and AI layoffs. Stripe is hiring a “Forward Deployed AI Accelerator” to embed within its marketing team and make “AI the default mode for all work.” Box is hiring an “AI Business Automation Engineer” to integrate AI agents across its cloud management platform. These roles exist to push employees who already have jobs to use AI more aggressively, which raises the question of what happens to the employees who do not adapt. GM’s decision this week to lay off 500 IT workers while simultaneously hiring for 250 AI positions is the clearest illustration of the dynamic: the same company is both creating and eliminating jobs in the same quarter.

The vibe coder is the newest category. The term, popularised by AI coding tools that allow non-engineers to build functional software through natural language prompts, has moved from internet slang to job listings. Lovable, a vibe-coding platform, is hiring professional vibe coders. TikTok is looking for a product designer who can create prototypes using “code and AI tools.” YouTube wants an “AI Solution Architect” who can “bypass traditional, slow-moving development cycles by utilizing AI-assisted development (vibe-coding) and low code solutions.” Engineering leaders are still figuring out how to measure the productivity gains from AI coding tools, but the job market is already pricing the skill as a standalone qualification. TikTok’s role starts at $108,000. YouTube’s starts at $149,000.

At the bottom of the AI jobs pyramid sit the gig workers who train the models. Companies like Scale AI and Mercor employ workers to evaluate creative writing output, train translation capabilities, and refine AI reasoning. Traditional gig platforms including Uber, DoorDash, and Instawork are also offering jobs that pay users for uploading photos and videos of chores and tasks that will be used to train AI systems. Depending on experience and task complexity, workers earn anywhere from $15 to roughly $200 per hour. The barrier to entry is lower than for any other AI role, but so is the security.

At the top sits the Chief AI Officer. PwC appointed one in July 2024. Accenture created a chief responsible AI officer the same year. Raymond James established a “Principal AI Architect” in 2025. Local governments are following: Arkansas is hiring a Chief AI Officer at a starting salary of just over $117,000. Glassdoor estimates private-sector pay for the role between $265,000 and $494,000.

The graduates entering this market are doing so at a moment when AI is simultaneously the most in-demand skill and the technology most frequently cited as the reason for layoffs. Detroit’s Big Three automakers have cut 20,000 white-collar jobs while posting 400 AI positions. Salesforce cut 4,000 support staff and is spending $300 million on Anthropic tokens. The pattern is consistent: the jobs AI creates pay more, require more specialised skills, and are fewer in number than the jobs it eliminates. The net effect on employment is a question economists will debate for years. What is not in debate is that the job titles on the name tags at the next networking event will look nothing like the ones from two years ago.