Running DOOM on weird obscure hardware is a fun hacker pastime that’s been around for a long time now. It’s always enjoyable to see someone port it to an egg timer, or a hat, or whatever else. But what about running the iconic shooteron a CPU of your very own? [Armaan] and [Liam] have done just that.
The CPU in question was designed at the logic gate level, deployed on to an FPGA, and hooked up with the necessary peripherals to run as a going concern. Early testing of the CPU involved running straightforward code to generate Mandelbrot sets and to play a simple game of Pong. But [Armaan] and [Liam] had bigger goals: to port the game that everybody ports to everything. Doing that took some work.
To get DOOM running, the CPU had to get faster, and it needed many tweaks to how memory was handled. There was also work to be done to create a keyboard interface, an HDMI video output, and a hardware timer. From there, the game itself had to then be ported to the custom CPU’s architecture. Eventually, the duo had the game running… at a glacial 0.7 FPS. A success, but not the magical end result that was desired. A bump to clock speed and further optimizations and compiler tweaks eventually got the game up to an impressive 15-20 FPS. The goal for future work is to push it to an entirely-playable figure of 30 FPS or better.
It’s worth checking out the (apparently unembeddable) videos on Instagram to see the CPU in action. We’ve also featured plenty of fun DOOM ports before, too. If you’re brewing up custom CPUs or DOOMports of your own, keep them coming to the tipsline. The latter in particular is often a wonderful milk run for the writer that happens across it. Happy hacking out there!
The UK government's blueprint for turning Britain into an AI superpower largely ignore the water datacenters need, prompting warnings that these facilities could end up competing with homes and businesses for supply. The warning comes as parts of southern, central, and eastern England move closer to drought, putting the pressure on water availability firmly on the agenda. MPs were told earlier this year that official forecasts of England's future water needs excluded datacenters, despite government efforts to encourage more of them to power AI development. In written evidence to a parliamentary committee, Water UK, the trade association for Britain's water industry, said the government's policies for delivering AI Growth Zones under the AI Opportunities Action Plan appear to assume the country will have ample supplies of H₂O. This, Water UK argues, limits the investment that water companies can plan for, including new reservoirs, leaving future supply constrained by forecasts that it says are already too low. The submission claims some housing projects are already being blocked and businesses barred from expanding because water isn't available. Water UK says supply is so constrained that the Environment Agency has effectively imposed a moratorium on business expansion in East Anglia and parts of Essex. Water UK is calling for new supply infrastructure to be fast-tracked and abstraction limits to be reviewed. It also wants operators to cover the infrastructure costs their facilities create, minimum efficiency standards for server farms, and requirements to use non-potable water wherever possible. Current economics push operators toward cooling methods that save energy rather than water. Where water is relatively cheap compared to energy, operators favour using more of it to use less power, as The Register previously detailed. UK energy prices are among the highest in the developed world. The submission says global datcenter water use is expected to more than double between 2025 and 2050, with services-based economies like the UK likely to account for a disproportionate share of that growth. T he UK datacenter pipeline is almost double that of the next-largest country in Europe, with ambitions to treble capacity by 2030. A government report [PDF] says most of the UK's current datacenter capacity is sited in and around London, with more than 1,000 MW located there. Europe's largest datacenter cluster is reportedly in Slough, Berkshire, home to as many as 35 facilities. The same report notes UK datacenter operators aren't required to report water use, so no official consumption figures exist, though the Water Research Centre estimates English facilities use almost 1.9 billion liters a year. According to PublicTechnology, the Government Digital Sustainability Alliance warned last year that AI is driving a significant, yet often underestimated, rise in water consumption, threatening water security both globally and in the UK A Water UK spokesperson told The Register: "We desperately want to supply the water that datacenters need. That's impossible because the UK government explicitly excludes them from its water planning framework for England." Water companies have been given approval to build ten reservoirs, but they will take years to complete, the spokesperson added. "We need planning hurdles cleared and water efficiency standards for datacenters introduced. Major businesses should pay for the water infrastructure they need so that the costs don't fall onto households." The Register also asked the relevant government department for comment, but had not received a response by the time of publication. ®
As technologies change and adapt, we’re often left with seemingly useless junk that has nowhere to go. Certainly anyone still sitting on a pile of floppy disks feels this way sometimes, but odds are anyone who owns a mining ASIC or an NFT can attest to that as well. The trillions of dollars flowing into GPU-based data centers will likely become the next victim of this trend, so if you want to capitalize on the losses of some venture capitalist you’ll want to figure out a way to get GPUs meant for a server into your desktop doing useful work.
Of course, calling these devices GPUs is a bit of a stretch compared to the Radeon and GeForce cards many of us are used to using for gaming. These don’t even have a PCIe slot or video output, after all. But, as [Oscar] notes, the VRAM and GPU cores are very real and can still do useful work. An adapter board is able to mate a Tesla V100 SXM2 16 GB GPU to a standard PCIe slot, which solves the first problem, but the major downside from there is that the cooling fan for this unit was literally deafeningly loud. At 82 dB it was about as loud as a lawnmower, which is fine in a server rack but not great in a bedroom. [Oscar] found a way to tamp down the fan speed, making it usable in a home.
Without video output, the utility of these cards mainly comes from adding VRAM and compute for tasks that benefit from parallel computing. Using tensor splitting, [Oscar] is running a local LLM with this card alongside his RTX 4080, providing 32 GB of VRAM on his NixOS system. With his benchmarking tests, the LLM sports impressive stats for a self-hosted model, ranking somewhere around Claude Sonnet 4.6. What’s even more impressive is that this is all done for around £200, and with the rate the various LLM companies are ratcheting up pricing could pay itself back very quickly. If trading off performance for cost is acceptable, though, it’s possible to run local models on much less powerful hardware as well.
The company said it is reducing its headcount by 20%, or about 630 staff, to "support a leaner, more focused operating model" as it focuses on its AI Work Platform.
L-R: Mantra CEO John Patrick Mullin, Docugami CEO Jean Paoli, and Inveniam CEO Patrick O’Meara. The companies are partnering to make DGML a standard for AI, with Docugami turning documents into data, Inveniam verifying it on a blockchain, and Mantra providing the chain.
Jean Paoli has spent his career making documents readable by machines — first as a co-creator of XML, then helping build the file formats behind Microsoft Office. Now his Kirkland, Wash.-based startup, Docugami, is open-sourcing the technology at the heart of its business, betting it can become a standard way to turn documents into data that people and AI agents can trust.
The company is releasing its technology, called DGML (short for Document Graph Markup Language), under Apache 2.0, a widely used open-source license, so other developers and companies can adopt it.
The idea is to turn it into a shared standard that no single company owns, much as XML became a common foundation across the tech industry.
The move reflects a shift in where the value is created in AI. Docugami until now has made its money selling software that turns unstructured documents into usable data. It’s betting now that there’s more value in proving that data is trustworthy instead.
How it works: Docugami is teaming up with Inveniam, a Detroit company whose software helps big investors keep tabs on the mountains of paperwork behind real estate and other hard-to-value assets. Inveniam will record a kind of digital fingerprint of each piece of DGML data on NVNM Chain, its blockchain built with Mantra, a crypto firm that Inveniam is acquiring.
That means, for example, that a single fact buried in a 200-page lease — such as the rental rate, a renewal option, or a default clause — can be verified on its own, without exposing the whole document. An investor, auditor, or AI agent can trace it to the page it came from.
To work with documents, AI systems usually convert them into a simpler format first. DGML enters a growing field of contenders in that regard, competing with the popular Markdown format and DocLang, a new open standard for AI-ready documents backed by IBM, Nvidia and Red Hat.
The business model: This is a big move for a company of Docugami’s size, taking the 30-person startup in a new direction. Paoli is handing the industry the technology his team spent years building, and pinning the company’s future on a larger idea.
The plan is to make money not from the format itself but from the value of the trusted data. Once a company converts its leases or loans into DGML and anchors the key numbers on the blockchain, investors, lenders and auditors can pay to draw on that verified data.
Docugami will share in the revenue through its partnership with Inveniam. The company also stands to collect a small fee each time a piece of data is recorded on the chain.
The company is giving away the DGML format and a working version of the software, but not everything. Paoli said the company is keeping some of its own technology private, including AI models it has fine-tuned to read documents, and could sell those or other tools to enterprises.
“The business model of everybody is changing. And if you know any company where it’s not true, you need to tell me, because I haven’t met them yet,” Paoli said in an interview.
Docugami has raised about $13 million to date, including a $10 million seed round in 2020 that drew the first investment in Grammarly’s history.
The partnership: Paoli met Patrick O’Meara, Inveniam’s CEO, a few months ago, through a former Microsoft colleague who had become one of O’Meara’s advisers. They quickly realized they had been working toward the same idea from different directions.
Inveniam, founded in 2017, helps big investors keep track of assets that are hard to value, like office towers, private loans and infrastructure. It monitors the documents behind those assets and flags changes as they happen, and its clients include some of the world’s largest sovereign wealth funds, according to O’Meara.
What it lacked was a consistent way to break those documents into verifiable pieces. That is what Docugami provides.
“We’re not putting the data itself on-chain, just a fingerprint of the document. Change one bit, one byte, one pixel, and the hash won’t match,” O’Meara said.
Paoli said the project uses the underlying blockchain, not the token.
“Crypto as an industry has gone through a lot of changes in the last 18 to 24 months, and it’s growing up in a lot of ways. This is a real use case with fundamental value, not just pure speculation,” Mantra’s Mullin said in an interview.
The result is a division of labor: Docugami turns documents into data, Inveniam verifies it and brings the customers, and Mantra provides the chain where the proof is recorded.
The DGML specification, sample documents and reference code are at dgml.io and on GitHub.
Editor’s note: This story was updated after publication to correct the name of a competing document format, DocLang, and to note that Inveniam’s blockchain is called NVNM Chain.
Aurum PropTech is buying Housing.com in a $47.5 million all-stock deal, but the harder test will be integrating its platforms and customer data securely.
A little over a month after the Department of Defense (DOD) bragged that nearly half of its 3.5 million employees were using AI at work, members of the Army’s Combat Capabilities Development Command (DEVCOM) received an email informing them that they were burning through tokens, and needed to limit use.
“Although the Army CIO announced in May 2026 that they were offering unlimited tokens, by mid-June the Army CIO pool was exhausted of tokens and had to re-establish limits,” the email reads. The email goes on to say that although the Army has chosen to renew token usage at “its current levels,” it’s unclear “if the Army CIO pool will be renewed after 1 Oct.”
The Army uses Ask Sage, a multimodal generative AI platform where users can run different large language models (LLMs), including Alphabet’s Gemini, Meta’s Llama, and OpenAI’s ChatGPT. “Apparently the whole Army burned through the whole year of tokens for just one service,” says an Army employee who spoke to WIRED anonymously because they were not authorized to speak to the press.
The advent of generative AI spurred an enormous and controversial datacenter building boom that has seen almost anyone who knows how to run a bit barn try to expand their business ASAP. Fujitsu, however, wants out. The Australian outpost of the Japanese giant’s business this week announced the sale of five datacenters down under. The company said selling the bit barns “enables us to further invest in the technology services where customer demand is growing fastest.” In Australia, that apparently means “helping organisations modernise critical systems, strengthen cyber resilience, adopt sovereign AI, and access the high-performance and quantum computing capabilities needed for their next phase of transformation.” Fujitsu said its datacenter business “is a strong platform, and its next phase will benefit from dedicated commercial ownership and investment.” That new owner, private equity outfit Next Capital, may have its work cut out for it because some of the bit barns it bought appeared to be rather modest. Fujitsu’s manifest of its Australian properties lists one facility capable of hosting 92MW worth of kit, another with 28MW capacity, plus bit barns that can host 10MW, 4.8MW, 3MW, or 2MW worth of kit. Keen-eyed readers will have noticed that the paragraph above mentions six datacenters and that earlier in this story we said Fujitsu is selling five. The Register understands Fujitsu has already disposed of the other one to another buyer. Whatever Next Capital bought, it will surely be aware that a modern rack filled with AI kit can require 500KW or more. Fujitsu Australia’s littlest datacenters therefore won’t help the private equity company to catch the AI wave unless it invests in upgrades – a process that might be cheaper than building new AI-ready datacenters from scratch and could also involve fewer regulatory complications than greenfield builds. Next Capital was quiet about its plans, but shared a local media report suggesting it’s spent AUD$200 million ($139.97/£104 million) to do the deal. The firm promised continuity for tenants. Fujitsu Australia’s services business won plenty of blue-chip and government clients, and The Register understands many are long-term residents of the offloaded datacenters. Next Capital can probably therefore bank on solid cashflow for months or years to come. Fujitsu sold its US datacenter business in 2023 and at the time hinted at divestments elsewhere. The company has also “absorbed” its Japanese public cloud and quit the mainframe business. The Japanese giant plans to return to the big iron business with machines built on the Monaka CPU which it hopes to deliver next year, and perhaps also get into the quantum computing biz. ®
As more and more of the technology that we paid for turns becomes a subscription, there’s slowly been a momentum shift in the open source world of building replacements for these intrusive rent-seekers. We see this all of the time for self-hosted media and communications servers, but now we’re starting to see it in hardware as well. The OOMWOO robotic vacuum cleaner is completely open source, from hardware to software, and requires no cloud services whatsoever.
Although it’s open source, not every component is something one could buy off the shelf. It does require a 3D printer for most of the parts, but assuming that requirement is met most of the rest of the build comes together easily enough. For compute it relies on a Raspberry Pi running ROS 2 software and is set up to integrate easily with other existing open tools and projects such as Home Assistant. Like its proprietary cousins it can sense and map the rooms its placed in, but this platform uses an inexpensive 2D lidar system to keep costs down.
Right now the project is not quite complete, so we’ll all have to keep our eyes on this one as the team building it progresses. But they do have most of the software development done and the bill-of-materials is in progress. As an open project it’s being developed by many volunteers and there are a lot of areas available to contribute to as well, all currently set up on the project’s GitHub page. Right now many of those areas of effort are adapting the 3D printer files to off-the-shelf parts.
Iran's Islamic Revolutionary Guard Corps (IRGC) claims it hit an AWS datacenter in Bahrain months after taking it offline for the first time, in a move it claimed as retaliation for a US attack on a nuclear plant that was under construction. The IRGC said in a statement on Tuesday that it had struck back at what it called the "child-killing US Army" by attacking Amazon infrastructure in Bahrain, claiming AWS's "central data infrastructure" had been "destroyed" after being hit by "several cruise missiles," according to Google Translate. The IRGC said the claimed strike was retaliation for what Iran described as a US attack on the under-construction Darkhovin nuclear facility. A look at the AWS Health Dashboard shows that there are definitely issues in Bahrain and the UAE, with issues in both regions being blamed on the US-Iranian conflict. While the status of AWS me-central-1 (UAE) is just “unable to reliably support customer applications,” me-south-1 (Bahrain) is said to be “currently unavailable.” The last update to the state of AWS services in Bahrain and the UAE in the open issues area of the Dashboard was on April 30, and the Service History tab shows that every single AWS service in Bahrain has been offline for months. Per our previous reporting, Iranian state-affiliated media claimed that strikes on AWS infrastructure in Bahrain and the UAE were deliberate, after US and Israeli attacks on Iran in late February. AWS-hosted providers including Snowflake and Red Hat subsequently urged affected customers to fail over or move workloads to other regions after the facilities were damaged. AWS waived all usage-related charges for March 2026 in its me-central-1 region in the UAE following the attacks on its infrastructure. Iran has also reportedly designated facilities associated with Google, IBM, Microsoft, Nvidia, Oracle, and Palantir as legitimate targets for retaliatory strikes, citing their alleged support for US military operations. The fact that services in Bahrain have been unavailable for months, and continue to be offline, makes it challenging for El Reg to confirm the legitimacy of Iranian claims it hit the site again. We’ve reached out to AWS to learn more, but didn’t hear back. ®
Linux-focused gaming site Boiling Steam breaks down the historical data, showing how the Deck's chart position has fallen from fifth place immediately after orders resumed in late May down to 14th place for two weeks in early July (the hardware sits at 12th place in the current edition of those charts).
That's in sharp contrast to 2025, when the Steam Deck never dropped below seventh place on the bestseller charts and was only rarely outside the Top 5. Then Valve warned of "intermittent shortages" for the Steam Deck starting in February, leading to weeks of lower chart positions before the Deck became completely unavailable until May.
With the new "Replace Audio" tool, users can update the music on their post at any time while keeping the post's existing likes, comments, shares, and reach intact.
From left: Clarify CEO Patrick Thompson, Seam AI CEO Nicholas Scavone, and Clarify CTO Ondrej Hrebicek. (Clarify and Seam Photos)
Clarify, the Seattle-based AI startup that has raised more than $22 million to take on Salesforce and other CRM incumbents, has made its first acquisition: San Francisco-based Seam AI.
Seam’s technology monitors buying signals across the web — such as funding rounds, hiring, website activity, and executive job moves — and surfaces them to sales teams. Clarify plans to fold the technology into a new product called Clarify Signals, slated to launch later this year.
Clarify is led by co-founders Patrick Thompson (CEO) and Ondrej Hrebicek (CTO), who previously co-founded Iteratively, a Seattle data-analytics startup that was acquired in 2021 by Amplitude, the publicly traded digital-analytics company.
Rationale: Clarify says the deal is part of a shift beyond what it calls a “system of record” that tracks what already happened to a “system of awareness” that flags what’s about to happen.
Thompson said the Seam deal fills a gap in what Clarify’s own AI can pull from the open web, giving the CRM access to proprietary datasets that can’t be reached with a simple search.
“The value that Seam is providing is typically the information that’s not necessarily easy to get from the web,” Thompson explained in an interview. “It’s the harder stuff to find.”
Hrebicek said Clarify’s customers have been looking for a bigger and richer dataset — the ability to “look around the corners on who would be a good lead.”
Deal points: Financial terms weren’t disclosed. Clarify, which had raised a total of $22.5 million in its seed and Series A rounds from investors including U.S. Venture Partners, Gradient Ventures, and Madrona, said it brought in additional funding as part of the deal but did not disclose the amount.
As part of the acquisition, five Seam employees are joining Clarify, including Seam co-founder and CEO Nicholas Scavone. With the deal, Clarify is adding a San Francisco office alongside its Seattle headquarters. The company now has 30 people total.
Backstory: Scavone started Seam in 2020 after five years at Okta, where he saw teams accumulate many different sales and marketing systems, with customer data scattered across all of them.
Seam raised $7 million including angel funding and a seed round led by Bessemer Venture Partners in April 2024. It counts Zapier, GoFundMe, Drata, and Betterment among its customers. Existing customers are on hold while the technology is integrated into Clarify, but many have already indicated they plan to move over to the new platform.
Scavone said he had been weighing whether to raise a new round or find a home for the company when he and Thompson, who have known each other for years, began talking about a combination.
“We’re all going after the same big incumbents here,” he said, explaining that he ultimately decided Seam had a better chance of taking on the market’s dominant players by joining forces with Clarify than as a standalone company.
In a post announcing the deal, the Seam and Clarify founders said they “realized we weren’t building competing products—we were building different halves of the same future.”
Landscape: Clarify is entering a crowded field. Sales-intelligence platforms like Clay, ZoomInfo, and Apollo already sell third-party data to revenue teams, and 6sense and Demandbase lead the account-based marketing category Seam had been targeting.
Thompson said one edge for Clarify is that signals arrive inside the CRM sellers already use, not a separate dashboard.
The company was co-founded in early 2024 by Thompson, Hrebicek, and Austin Hay, a marketing-technology operator who served as co-CEO alongside Thompson. Hay departed in September 2025 and is now with Khosla Ventures, per his LinkedIn.
What’s next: Clarify plans to launch Signals later this year, Thompson said, noting that the company is considering raising additional funds in a Series B round early next year.
Gemini Notebook’s new Collections help users organize growing research libraries as Google expands the former NotebookLM across Gemini and Search, although the feature remains more limited than a proper folder system.
Login
