CoreWeave IPO Signals AI Market Shift; New AI Tools Enhance Communication and Security

[00:00:02] It's Friday, March 28th, 2025 and I'm Dave Solt for Things to Know Today. The CoreWeave IPO taps the brakes on AI hype. Still, it pulls in $1.5 billion backing for GPU-powered growth. AI agents start talking, retail shelves get smarter, and accents get a makeover. Here comes the new AI stack. Microsoft still making money for everyone, but Windows 10's exit could cost small businesses big.

[00:00:30] And, from hotel tasks to missile warnings, who's watching the AI? A deep dive into control and risk. This is the Business of Tech. CoreWeave priced its initial public offering at $40 per share, raising $1.5 billion, marking the largest tech offering in the U.S. since 2021. The company, which provides access to NVIDIA graphics processing units for artificial intelligence workloads,

[00:00:58] initially aimed for a share price between $47 and $45. With this pricing, CoreWeave's valuation is estimated to be about $19 billion. Shares are set to begin trading on the NASDAQ under the ticker sign C-R-W-V. This is a reduction of 23.5% from the originally planned valuation. Why do we care? CoreWeave's IPO pricing is a significant downshift from its original target range,

[00:01:27] resulting in that 23.5% haircut in expected valuation. While it still represents the largest U.S. tech IPO in years, the reduced pricing is a clear signal that investor sentiment around AI infrastructure plays may be shifting, and not necessarily in a bullish direction. While the markdown is notable, this is still a $1.5 billion raise at a $19 billion valuation, partly a sign of collapse.

[00:01:53] The size of the offering signals that public markets are still willing to back AI-focused infrastructure, albeit with measured optimism. And let's not forget, CoreWeave is capitalizing on a real bottleneck in AI development, GPU scarcity. Its tight partnership with NVIDIA gives its access to critical hardware, and demand is still high among AI model developers, particularly those priced out of hyperscaler contracts. This IPO repricing doesn't signal the bursting of the AI bubble outright,

[00:02:23] but it is a firm tap on the brakes. The path to profitability, the durability of AI workload demand, and the reliance on NVIDIA's supply chain dominance are all under sharper scrutiny. For ITC service providers and infrastructure players, the message is this. The demand for AI support services is real, but frothy valuations and easy capital are not a given.

[00:02:47] With every new breach and threat that I cover, it's clear that cybersecurity isn't a luxury anymore. It's a necessity. That's where Huntress comes in. Their fully managed cybersecurity platform is built for every kind of business, not just the 1%. Huntress seamlessly integrates their products and threat hunting team. Their EDR, ITDR, SIM, and security awareness training solutions are purposely built

[00:03:12] for their elite 24x7 security operations center to stop threats before anyone else even spots them. This potent combination of purpose-built cybersecurity and threat hunting expertise is one of the many reasons why G2 users have voted Huntress the number one rated EDR for growing businesses. To see what people-powered cybersecurity looks like, visit Huntress.com slash MSP Radio.

[00:03:41] PwC has launched a new platform called AgentOS, designed to facilitate communication between artificial intelligence agents within enterprises. This initiative comes after the company spent 19 months developing and deploying AI agents for various clients, during which they observe that these agents often operate independently, like ships passing in the night. PwC aims to transform this dynamic by creating a collaborative environment for agents, allowing them to work together effectively.

[00:04:10] The platform integrates with major systems, including those from Anthropic, Google Cloud, and Microsoft Azure, enabling companies to automate complex tasks. Once implemented, clients can deploy multi-agent solutions within just a couple of weeks. With over 250 internal agents developed, PwC emphasizes the importance of interoperability among AI agents to enhance their effectiveness and utility in the workplace. Intel announced three new initiatives aimed at accelerating

[00:04:38] the adoption of artificial intelligence at the edge, targeting industries such as retail, manufacturing, and smart cities. The company introduced Intel Edge Systems, Edge AI Suites, and the open-edge platform to address the unique challenges posed by over 100,000 real-world edge implementations. These initiatives are designed to help organizations integrate AI into their existing infrastructure while achieving cost and performance goals. Dan Rodriguez, Intel's corporate vice president,

[00:05:07] emphasized the importance of expanding AI use in low-power, cost-sensitive environments. The new offerings include standardized blueprints for hardware integration, industry-specific software development kits, and a modular open-source platform for managing edge and AI applications efficiently.

[00:05:33] I also noted two interesting language-related products. A new AI tool by CRISP, known for its noise cancellation and transcription services, can convert a speaker's accent to American English in real time. The tool currently supports 17 Indian dialects and aims to help native English speakers better understand non-native speakers without altering their natural voice. Tested in call center enterprises, the AI was trained on thousands of speech samples

[00:06:03] and is compatible with popular meeting platforms like Zoom and Microsoft Teams. An off-Broadway theater in New York is using AI-powered live translations to enhance accessibility for non-English-speaking audiences. The longest-running play in New York, Perfect Crime, allows attendees to scan a QR code for translations in over 60 languages, with around 25 to 30 users benefiting from this service during eight shows each week.

[00:06:31] The translation technology developed by Silicon Valley startup Wordly promises high accuracy and aims to expand its language offerings in the future. Why do we care? PwC launching a platform for interagent communication is noteworthy not just because of the technology,

[00:06:54] but because of what it admits, that many enterprise AI deployments today are isolated, one-off tools that don't interact well. The metaphor of ships passing in the night underscores the current fragmentation across internal AI efforts. AI agents lack standard protocols. PwC's platform might help internally, but without cross-enterprise standards, agent interoperability risks staying vendor-specific.

[00:07:21] This could lead to ecosystem lock-in rather than broad interoperability. CRISP's real-time accent conversion tool and Wordly's live theater translation system show AI tackling long-standing accessibility and communication challenges in new ways. I do believe there's real value here. The through line here is infrastructure-level evolution, whether it's coordinating AI agents, scaling edge deployments, or enabling inclusive communication.

[00:07:49] Microsoft recently celebrated its 50th anniversary and put out a blog post with some useful data. The company boasts a partner ecosystem of over 500,000, emphasizing the collaborative spirit that has driven innovation from personal computing to cloud technology and now artificial intelligence.

[00:08:07] According to a survey by IDC, for every $1 of Microsoft revenue, services partners earn $8 and software partners earn nearly $11, highlighting the substantial economic opportunities available. In this story, it was almost a big idea. From Dark Reading, as the retirement date for Windows 10 approaches, small and medium-sized business IT managers are either implementing or planning their operating system migration strategies.

[00:08:34] The deadline for support is set for October 2025, and while the financial implications of hardware upgrades are widely discussed, hidden risks such as legacy systems and misconfigurations may expose organizations to vulnerabilities. Forrester Senior Analyst, Patty Harrington emphasizes the necessity of asset management in securing enterprise data, stating that organizations need to know what devices are connected to their networks.

[00:09:00] Brian Marlott, Chief Regional Officer at Zycel, notes that many companies lack asset management software, relying instead on vulnerability management tools to identify and manage their hardware. Critical steps for businesses to mitigate risks include isolating legacy systems, employing endpoint detection software, and tightening governance to prevent the use of outdated operating systems in internet-facing environments.

[00:09:25] Additionally, as cyber insurance policies require businesses to address vulnerabilities from unsupported products, organizations must explain their strategies to insurers to avoid potential policy invalidation. Now why do we care? The anniversary blog offers more than nostalgia. The updated IEC data shows that Microsoft's economic engine continues to generate outsized value for its partners.

[00:09:50] At $8 in services revenue for every $1 Microsoft earns, and $11 in software revenue for ISVs per $1 of Microsoft's earnings. The economics still work. These multipliers validate that Microsoft remains the most predictable platform for services firms to build around. Whether you're delivering integration, migration governance, or application modernization. That said, these IDC estimates are aggregate and directional.

[00:10:17] They don't account for margin compression, rising customer acquisition costs, or the increasing automation of previously billable services. The opportunity is there, but low-value partners may see diminishing returns. The real story in the Windows 10 phase-out isn't about replacing outdated laptops. It's about the messy, risky underbelly of unmanaged infrastructure. Most coverage is focused on upgrade costs. But the larger problem is visibility and control.

[00:10:47] This is a major services opportunity. SMBs don't have the internal muscle to handle this transition securely. There's a clear demand for asset discovery, segmentation, endpoint protection, and compliance mapping services. And time for some big ideas. In a recent article in MIT Technology Review, experts from Hugging Face warn that surrendering control to autonomous AI agents could lead to significant risks.

[00:11:13] While these systems, designed to perform complex tasks without direct human oversight, promised to enhance our lives, they may also introduce vulnerabilities, such as unauthorized transactions and privacy breaches. The researchers emphasize that as AI agents become more flexible, human control diminishes, raising concerns about safety and security. Historical examples like the false alarm about Soviet missiles in 1980 highlights the need for human oversight in decision-making processes.

[00:11:43] The article advocates for the development of open-source AI systems, which would allow for greater transparency and control, ensuring that technology serves human interests rather than undermines them. Over on Runtime, Writer CEO May Habib discusses the challenges of integrating generative AI into existing applications, emphasizing the shift in focus from model builders to development tool vendors as businesses seek practical applications.

[00:12:10] Writer has raised $326 million to build large language models and software tools to assist companies in app development, having the necessity of support for enterprises navigating the complexity of agenting AI. Khabib noted that while some automations in industries like hotel management are easier to implement, areas requiring human intervention, such as customer support, pose significant challenges due to the lack of standard procedures.

[00:12:36] And also on Runtime, an exploration of the growing excitement around the Model Context Protocol, or MCP, introduced by Anthropic last November as an open standard for building secure connections between data sources and AI tools. As major companies like Microsoft and Cloudflare adopt MCP, the protocol promises to simplify the development of AI agents, akin to how application programming interfaces revolutionized web-based computing.

[00:13:03] However, experts warn that MCP is not yet ready for widespread enterprise use as it lacks built-in permissions models and has unresolved security concerns regarding payment processes. Why do we care? Let me give you some questions to ponder. What organizational or regulatory mechanisms can ensure fail-safe controls across independently operating AI agents? This is a service that customers embrace.

[00:13:31] How do IT service providers differentiate in a world where generative AI platforms increasingly abstract away the hard stuff? And what kind of observability and trust signals will enterprises and organizations require before exposing internal systems to agent-based interactions via MCP? This episode is supported by Flexpoint.

[00:13:56] Flexpoint offers a purpose-built payment solution from managed service providers, automating billing operations to enhance efficiency and cash flow. With features like accounts receivable automation, branded client portals, and secure same-day payments, Flexpoint streamlines financial management. Integrations with accounting software such as QuickBooks and Xero, as well as professional services automation tools like ConnectWise and Autotask, ensure seamless data synchronization.

[00:14:22] Experience improved cash flow and client satisfaction with Flexpoint's comprehensive platform. Learn more at GetFlexpoint.com slash MSP dash radio. Thanks for listening. Today is National Hot Tub Day, National Something on a Stick Day, and National Black Forest Cake Day. It's also Respect Your Cat Day, apparently. Nerdy Ocon will be held in Palm Springs, California from April 7th through 9th.

[00:14:53] Visit NerdyOcon.com to learn all about it. The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines posted at BusinessOf.Tech. If you've enjoyed the show, make sure you've subscribed or followed on your favorite platform. It's free and helps directly. Give us a review, too. If you want to support the show, visit Patreon.com slash MSP Radio, and you'll get access to content early.

[00:15:21] Or buy our Why Do We Care merch at BusinessOf.Tech. Have a question you want answered? We take listener questions, send them in, ideally as a voice memo or video to question at MSP Radio.com. I answer listener questions live on our Wednesday live show on YouTube and LinkedIn. If you've got a comment or a thought on a story, put it in the comments if you're on YouTube, or reach out on LinkedIn if you're listening to the podcast.

[00:15:47] And if you want to advertise on the show, visit MSP Radio.com slash engage. Once again, thanks for listening, and I will talk to you again on our next episode. Part of the MSP Radio Network.