Full-time, in-person in the London office.
Perplexity AI provides an answer engine that delivers precise, reliable responses to user questions by using up-to-date sources. It serves individuals who want quick answers and businesses that need detailed information, drawing on current data and source links to back its results. The product works by retrieving information from reputable sources, compiling concise answers, and presenting citations to ensure trustworthiness. It differentiates itself from competitors by prioritizing current, source-backed information and offering access that spans personal and enterprise use, potentially supported by subscriptions, advertising, and partnerships. The company's goal is to be a dependable tool for immediate, accurate information across a wide range of queries.
Company Size
1,001-5,000
Company Stage
Private
Total Funding
$1.8B
Headquarters
San Francisco, California
Founded
2022
See people who can refer or advise you
Help us improve and share your feedback! Did you find this helpful?
Health Insurance
Dental Insurance
Vision Insurance
401(k) Retirement Plan
Company Equity
Revolut chatbot Rita charges users a break fee it had already waived. Revolut chatbot Rita mischarged customers seeking to cancel their premium subscriptions, imposing a break fee the digital bank had already agreed to waive following a reduction in a key membership perk. The automated assistant, formally named Revolut's Intelligence Troubleshooting Assistant, told at least one customer they owed £55 to downgrade from the company's Ultra plan (the same amount as a single month's subscription) despite Revolut having quietly extended fee-free cancellations to affected members. WeWork cuts trigger Ultra subscription revolt. The incident follows Revolut's decision to reduce the number of WeWork office locations available to Metal and Ultra subscribers. WeWork cited economic factors for the reduction. Around 265 sites remained accessible to Revolut subscribers, but more than half are in the US, where Revolut's customer base is considerably smaller than in Europe. Paris, Milan and Edinburgh lost all of their WeWork locations. According to Allwork.space, Brussels, Dublin and Prague were also stripped of all sites, cities absent from Revolut's initial communications on the change. Customers were given no advance notice. Revolut's general terms and conditions commit the bank to 30 days' notice before altering an existing product or service. A person close to the bank said partnership perks are governed by separate terms, which permit both Revolut and the partner to alter or cancel benefits at any time, and that participating locations are also controlled by WeWork. Revolut's own paid plan terms state that the provision, scope and terms of subscription benefits may be changed or cancelled by Revolut or a partner at any time, and that advance notice of detrimental changes will be given wherever possible. How Revolut chatbot Rita got it wrong. In an exchange seen by City AM, Revolut chatbot Rita acknowledged the customer's complaint but still demanded payment. Rita told the customer: 'I understand your frustration regarding the changes to the Wework and Perplexity benefits, and that this break fee is unexpected given the situation.' Revolut had, in fact, already decided not to charge the fee. A spokesperson said: 'We are not charging an early termination fee for customers who reach out to customer support wishing to downgrade their plan, as a result of this recent update.' The spokesperson added: 'We apologise that this option was temporarily unavailable to some users.' The bank also confirmed it was replacing Perplexity with ChatGPT in its plan bundle from September, having informed customers at the start of August. TechTimes reported that Ultra and Metal subscribers will receive 12 months of ChatGPT Go free. The switch marks Revolut's third AI provider partnership in under a year, following Perplexity and the Swedish platform Lovable. Subscriptions drive Revolut's record revenue. The episode touches a commercially sensitive line. Subscriptions are among Revolut's fastest-growing revenue streams, with turnover reaching $936m (£708m) in the latest financial year, a 67% increase from $541m (£423m) in the prior year, according to the Revolut 2025 Annual Report. Total group revenues rose 46% to $6.0bn (£4.5bn), up from $4.0bn (£3.1bn) in the prior year. Profit before tax climbed 57% to $2.3bn (£1.7bn), the company's 2025 results show, making it a fifth consecutive year of net profitability. The customer base underpinning those figures stood at 52.5 million retail customers at end-2024, up 38% from 38.0 million a year earlier, per the Revolut 2024 Annual Report. Retaining premium subscribers at £55 a month is central to maintaining that revenue trajectory. Revolut said it retains a good working relationship with WeWork, describing the co-working firm as a valued partner. WeWork's own relationship with Revolut extends beyond perks: the two have a commercial tenancy arrangement, with Revolut taking space for staff at a WeWork site in Dallas. Whether Rita has been corrected is the immediate test. Any repeat of automated misbilling on a product line generating nearly a billion dollars a year would be a harder apology to script.
HP targets enterprise AI workloads with new ZBook Ultra G3a mobile workstation. 3 minutes read HP has unveiled the HP ZBook Ultra G3a 16, a mobile workstation designed to bring large-scale local AI processing and agentic AI capabilities to professional environments. The new workstation combines AMD-powered processing with Perplexity Computer and support for professional applications, allowing organisations to run demanding AI workloads locally rather than relying exclusively on cloud infrastructure. HP says the ZBook Ultra G3a can support LLMs with up to 300 billion parameters locally, with configurations offering up to 192GB of unified memory, including up to 160GB of dedicated VRAM. The company is positioning the device as a personal AI platform for professionals as organisations move from AI experimentation towards more practical and production-oriented applications. "Professionals need AI experiences that are powerful, reliable, and integrated into the way they work," said Jim Nottingham, Senior Vice President and Division President, Advanced Compute and Solutions, HP Inc. "With the HP ZBook Ultra G3a 16", HP is bringing a local AI factory and workflow-ready tools together to help customers apply agentic AI to their most demanding work." Bringing large AI models closer to professionals The ZBook Ultra G3a has been co-engineered by HP and AMD and is powered by up to the AMD Ryzen AI Max+ PRO 495. Its unified memory architecture is designed to support large AI models, AI agents and professional applications simultaneously. "Together, HP and AMD are ushering in the era of Personal AI, where extraordinary AI power moves from the data center into the hands of professionals. For years, the most powerful AI has lived in the cloud," said Jack Huynh, Senior Vice President and General Manager of the Computing and Graphics Group at AMD. "Now, with up to 192GB of unified memory and the ability to run models up to 300 billion parameters locally, that power is right beside you, working alongside you and your most demanding applications. The new ZBook shows what's possible when AI becomes truly personal, giving people the freedom to explore bigger ideas, solve harder problems, and bring their most ambitious ideas to life." For organisations, local AI processing can offer an alternative to sending every workload to cloud infrastructure, potentially providing greater control over data, latency, privacy and inference costs. Agentic AI integrated into professional workflows Perplexity Computer comes preloaded on the ZBook Ultra G3a, providing AI capabilities designed to support defined workflows and tasks. "Professionals shouldn't have to build their own AI stack to put their workstation to work," said Ryan Foutty, Vice President of Business at Perplexity. "With HP, we're bringing them Portable Computer, a mode of Personal Computer that runs AI models, agents, and tools on the device. Work starts locally, and when a step needs frontier intelligence, users can choose to bring in a cloud model without starting over. Businesses get the benefits of local AI without giving up access to the cloud or control over when they use it." The integration with Autodesk Revit is enabled through Perplexity Portable Computer for Windows's MCP Connector, allowing local AI capabilities to work alongside project information on the device. HP and Perplexity have also co-developed task-based Skills for HP customers, intended to help automate repetitive work and reduce the amount of manual configuration required to use AI within professional workflows. Mobile form factor with workstation performance The ZBook Ultra G3a 16 is designed for professionals who need workstation-class capabilities away from a fixed desk. It starts at 4.2 pounds and measures 17.9mm thick. HP's thermal design supports up to 100W of performance for advanced workloads, with the company reporting an 81.8% increase in TDP compared with the previous design. The system is designed to support workloads involving Autodesk Revit 2027, Autodesk 3ds Max and Chaos Enscape, alongside AI applications. Connectivity options include secure 5G through HP Go, while an IT-replaceable keyboard and other serviceability features are intended to simplify maintenance and extend device life. Expanding HP's local AI portfolio The ZBook Ultra G3a forms part of HP's broader agentic AI portfolio, which also includes the HP ZGX Nano, HP ZGX Fury, HP Boost and HP Remote System Controller. HP is also offering a Local AI Value Calculator, designed to help organisations assess the potential financial and performance impact of moving selected AI workloads from cloud infrastructure to local workstations. The company says its approach provides organisations with greater flexibility around AI inference, memory capacity, scaling and device lifecycle management as AI models and workflows continue to evolve. The HP ZBook Ultra G3a 16 is expected to become available in October 2026. Pricing will be announced closer to availability.
Crusoe has signed a multiyear cloud agreement to supply Perplexity with Nvidia AI chips for model training and inference, Bloomberg reported Tuesday. The deal's financial terms and capacity were not disclosed. Under the agreement, Crusoe will manage the technology throughout its operating life. More than 1,800 Crusoe employees will receive access to Perplexity's Enterprise Max product, which provides AI-powered search tools. The contract adds to Crusoe's growing cloud business, which recently secured a five-year agreement with Jane Street worth roughly $13 billion. In June, Crusoe reported 4.9 GW of contracted AI infrastructure capacity across its data centre projects and cloud platform, with a development pipeline exceeding 40 GW. Crusoe opened a 400,000-square-foot manufacturing plant in Tulsa, Oklahoma, on Monday.
Crusoe signs multiyear cloud deal with Perplexity for AI models. The AI infrastructure provider will host and manage Nvidia chips powering Perplexity's search models, adding to a growing roster of major cloud contracts an hour ago Sponsored: Vera - AI-powered prediction market intelligence, built for serious analysts Explore Vera Crusoe, the data center and AI cloud infrastructure company, has locked in a multiyear cloud capacity agreement with Perplexity AI to support the search startup's model training and inference workloads. The deal puts Crusoe in charge of hosting and managing Nvidia AI chips on Perplexity's behalf, a managed GPU arrangement that gives the AI search company another option beyond the hyperscaler giants it already relies on. For Crusoe, the partnership is another headline-grabbing contract in what's been a breakout stretch. Earlier in September, the company announced an approximately $13 billion five-year AI cloud contract with Jane Street, the quantitative trading firm. What the deal actually covers. Under the agreement, Crusoe will host Nvidia chips dedicated to Perplexity's AI operations and manage those chips over their full lifetime. Crusoe SVP Erwan Menard confirmed the company's role in overseeing the hardware on an ongoing basis, which means Perplexity is essentially outsourcing not just the compute but the operational headaches that come with running large GPU clusters. The deal also includes a reciprocal element: Perplexity will make its Enterprise Max product available to more than 1,800 Crusoe employees. AI, tech, and the markets they move - in one daily briefing. Daily. Free. Join 34,000+ readers across crypto, finance, and policy. Financial specifics like chip volume, total contract value, and precise timelines weren't disclosed. Perplexity's cloud diversification strategy. Perplexity has been deliberate about not putting all its eggs in one cloud basket. The company already has substantial commitments with AWS and struck a $750 million deal with Microsoft Azure reported back in January 2026. Adding Crusoe to the mix gives Perplexity a third major infrastructure partner, one that specializes in AI workloads rather than general-purpose cloud computing. Crusoe's rapid ascent in AI infrastructure. The firm has built out what it describes as nearly 5 GW of AI infrastructure capacity, with a pipeline exceeding 40 GW. The company originally gained attention for using stranded natural gas at oil wells to power Bitcoin mining operations, converting what would otherwise be flared emissions into productive compute. Disclosure: This article was edited by Editorial Team. For more information on how Crypto Briefing create and review content, see its Editorial Policy.
You can now run Perplexity's local AI agent on your Windows PC. Your gaming GPU can now moonlight as an AI agent. Sep 15, 2026 - 8:11 AM ET Karandeep Singh / Android Authority Add Android Authority on Google: * Perplexity's Portable Computer can now run on Windows PCs with RTX GPUs. * It runs AI tasks locally, keeping sensitive files secure and saving token costs. * Portable Computer requires an RTX GPU with at least 24GB of VRAM. Perplexity has been rolling out feature after feature this year. After the agentic "Personal Computer," the company launched a new Portable Computer late last month for NVIDIA DGX Spark systems. Now, Perplexity and NVIDIA are finally making Portable Computer accessible to even more people. NVIDIA announced today that Perplexity's Portable Computer now works on Windows machines with compatible NVIDIA RTX GPUs. This means users no longer need a $5,000 NVIDIA DGX Spark setup to run Portable Computer. It also finally opens up access beyond DGX Spark and Linux-based systems. Perplexity Portable Computer runs completely on-device, with the AI models, tools, and the agent itself installed directly on the user's computer. This is useful for two main reasons: First, users can hand off tasks to the agentic AI without worrying about privacy because everything runs locally. Second, on-device tasks do not incur token costs, making them more cost-effective. For more advanced workflows, Portable Computer can still use cloud-based models. However, it asks for permission before offloading any tasks to more powerful cloud alternatives. Though Portable Computer supports Qwen 3.8 27B and a post-trained version of Qwen 3.8 called PPLX 27B, these models remain exclusive to NVIDIA DGX Spark systems. Users on Linux or Windows machines with RTX GPUs can only use Qwen 3.8 locally. However, Perplexity is working to add support for the NVIDIA Nemotron 3.5 Lightning model on both Linux and Windows. Portable Computer also requires an RTX GPU with at least 24GB of VRAM. That means you'll need an RTX 3090, 3090 Ti, 4090, or 5090 to run it, or a professional workstation GPU like the RTX Pro 4000, RTX A5000, or RTX Pro 6000. None of these GPUs are affordable, but they are considerably cheaper than a DGX Spark setup.