The Answer Engine AI News · Sydney

SpaceX has secured a $40 billion financing arrangement with Apollo Global Management specifically to purchase GPUs — the powerful chips that run AI systems. This deal gives SpaceX a way to fund its growing artificial intelligence projects without going through traditional bank loans or debt markets. The sheer scale of the financing reflects how expensive AI infrastructure has become: cutting-edge GPUs from companies like Nvidia can cost tens of thousands of dollars each, and building competitive AI systems requires thousands of them running in data centres. SpaceX, best known for rockets and Starlink satellites, has been quietly expanding into AI development. This financing structure is notable because it's collateralised specifically against hardware purchases rather than general corporate debt. For context, $40 billion is an enormous sum — roughly equivalent to building multiple large data centres from scratch. The deal signals that major infrastructure investors like Apollo see AI computing power as a safe, tangible asset worth backing with serious capital, and that even cash-rich companies like SpaceX are seeking creative financing to compete in the AI arms race without tying up their own capital reserves.

Anthropic, the company behind Claude AI, has filed paperwork to go public on the stock exchange. The filing reveals the company's revenue is climbing rapidly, but it's also losing significant money as it invests heavily in development. The documents show major businesses are adopting Claude for their operations, suggesting enterprise AI is moving beyond experimentation. What's unusual is Anthropic's governance structure: it's designed to let founders keep control and protect the company's AI safety mission even after going public. This matters because Anthropic is one of the main players challenging OpenAI's ChatGPT and is a key provider of the AI models that power many answer engines and AI assistants Australian businesses interact with daily. The IPO will give investors their first detailed look at the economics of running a major AI company—how much it costs to build these systems and whether the business model actually works at scale.

Anthropic, the AI company behind Claude, has filed paperwork to go public. The 335-page IPO prospectus reveals how the company sees itself and the AI technology it's building—and roughly 100 pages of that document are devoted to listing risks. This is significant because Anthropic has positioned itself as the "safety-focused" alternative to OpenAI and other AI labs, emphasising responsible development. The filing gives the public its first detailed look at Anthropic's financials, business model, and internal thinking. It comes after weeks of speculation about whether the company would pursue an IPO or remain private. The sheer volume of risk disclosures suggests Anthropic is being unusually candid about the uncertainties and potential downsides of AI development—a stance that aligns with its public messaging about taking AI safety seriously. For context, Anthropic was founded by former OpenAI executives who left over disagreements about the company's direction and partnership with Microsoft.

MIT Technology Review highlights an opinion piece arguing that the AI safety conversation is too focused on catastrophic, future scenarios while ignoring the real harms AI systems are causing right now. The piece, from youth advocacy leaders Sam Hiner and Emma Leiken, points out that everyday AI applications are already creating tangible problems for ordinary people—issues that don't make headlines but affect lives today. While tech leaders and policymakers debate theoretical existential risks from advanced AI, current systems are perpetuating bias in hiring, creating unfair outcomes in credit decisions, amplifying misinformation, and affecting people's access to services. The authors argue that young people, who will live longest with AI's consequences, are particularly concerned about these immediate, practical harms rather than speculative doomsday scenarios. The message is clear: the AI safety debate needs rebalancing. Addressing present-day algorithmic discrimination, privacy violations, and other concrete problems deserves at least as much attention and resources as preparing for hypothetical future catastrophes.

Tech companies are celebrating minimal safety improvements—like preventing chatbots from encouraging suicide or making it harder for kids to buy drugs on social media—as major wins. But these basic protections should have been built in from the start. The problem is that while companies tick these boxes, they continue releasing products that cause widespread, less dramatic harm every day. The current policy debate focuses almost entirely on catastrophic AI risks: rogue superintelligence, mass job losses, or election manipulation. This focus on disaster scenarios lets companies off the hook for the mundane damage their products already cause—amplifying misinformation, eroding privacy, manipulating behaviour, and harming mental health. These everyday harms affect millions of people right now, but they don't generate the same headlines or regulatory urgency as hypothetical future catastrophes. The article argues we need to rebalance our attention: yes, plan for worst-case scenarios, but don't let that distract from holding tech companies accountable for the real damage their products are causing today.

New research has found that ChatGPT often fails to mention certain brands in its responses, even when those brands appear in the sources it cites. The study compared logged-out ChatGPT responses and discovered that brands which didn't show up when ChatGPT searched for information were also less likely to be named in answers—despite being covered on the pages ChatGPT referenced. This means a brand could be featured in an article that ChatGPT uses as a source, yet still not get mentioned in the actual answer given to users. The issue appears linked to whether brands surface in ChatGPT's initial search process. This creates a visibility gap where businesses miss out on being recommended or discussed, even when they have legitimate coverage in authoritative sources. For businesses relying on content marketing and PR to build visibility, this research suggests a significant blind spot: traditional "get mentioned in quality publications" strategies may not translate into AI-generated answers, which are increasingly how people find information and make decisions.

Traditional click-through metrics don't work when AI search engines like ChatGPT, Perplexity, and Google's AI Overviews answer questions directly without sending users to websites. Tom Capper from STAT Search Analytics recommends three practical approaches to measure your visibility in AI search results. First, track brand presence by monitoring how often your business appears in AI-generated answers, even without clicks. Second, investigate citations to understand which sources AI engines are pulling from and whether your content is being referenced. Third, improve your attribution story by connecting AI visibility to downstream business outcomes like brand searches, direct traffic, or conversions that happen later in the customer journey. The core challenge is that AI search fundamentally breaks the click-based measurement model that's guided SEO for decades. Businesses need new frameworks to understand whether their content strategy is working when AI tools increasingly act as intermediaries between searchers and websites, summarising information rather than directing traffic.

Anthropic, the company behind Claude AI, has reported massive revenue growth but equally massive losses as it prepares for a potential public listing. While the company's sales are climbing rapidly—driven by businesses paying for access to its AI models—it's burning through cash at an extraordinary rate to fund the enormous computing power and infrastructure required to train and run advanced AI systems. The widening net loss highlights a central tension in the AI industry: these companies can attract customers and revenue, but the costs of development and operations are staggering. An IPO would put Anthropic's business model to the ultimate test, asking public market investors whether they believe the path to profitability is realistic or if current AI economics are unsustainable. The outcome matters beyond Anthropic itself—it will signal whether investors have appetite for AI companies with sky-high costs, and could influence funding and valuations across the entire sector. For now, the results show AI companies can sell their products, but questions remain about whether they can ever make money doing so.

iTnews published a photo gallery from a recent roundtable discussion on "shadow AI" security held at Mr Morris in Auckland. The event was hosted by The Instillery, a technology consultancy. The images document what appears to be a business-focused conversation about the security risks posed by shadow AI — the term for AI tools employees use without IT department approval or oversight. While the source material consists only of photographs from the event rather than detailed content from the discussion itself, the topic signals growing concern among New Zealand businesses about unauthorized AI tool adoption in their workplaces. Shadow AI creates potential data security, privacy, and compliance risks when staff use consumer AI platforms like ChatGPT or similar tools to process sensitive business information without proper governance frameworks. The roundtable format suggests this is becoming a priority issue for businesses seeking peer insights on managing AI adoption while maintaining security standards.