Core Viewpoint - The emergence of the "OpenAI Mafia" is reshaping the AI landscape, with former OpenAI employees founding numerous startups that have collectively raised significant capital and achieved high valuations, indicating a strong market demand for AI expertise and innovation [3][4][5]. Group 1: Company Developments - Mira Murati's new venture, Thinking Machines Lab, is launching with a valuation of $9 billion and aims to create AI solutions tailored to individual needs, although specific product details remain undisclosed [1][2]. - The "OpenAI Mafia" has already established over 30 startups since 2018, raising more than $9 billion in total funding, creating a comprehensive ecosystem that includes AI safety, infrastructure, and vertical applications [3][4]. - Notable companies include Anthropic, valued at $61.5 billion, and Ilya Sutskever's SSI, valued at $30 billion, showcasing the rapid growth and high valuations of these new entities [4][5]. Group 2: Talent Migration and Impact - A significant portion of OpenAI's talent has left, with over 45% of core technical staff opting to start their own companies, leading to a "talent explosion" that is influencing the entire Silicon Valley and potentially reshaping the global AI power structure [5][9]. - The departure of key figures, including Ilya Sutskever and other executives, has left OpenAI's core research team fragmented, with only two of the original eleven co-founders remaining [7][20]. - The trend of talent leaving OpenAI mirrors the historical "PayPal Mafia," which led to the creation of several successful companies, indicating a similar potential for the "OpenAI Mafia" [3][4]. Group 3: Market Dynamics and Competition - OpenAI is facing increasing competition from newer entrants like DeepSeek, which has developed models that rival ChatGPT at a significantly lower training cost, threatening OpenAI's market position [18][19]. - The internal conflicts and cultural shifts within OpenAI, including a focus on commercialization over exploratory research, have contributed to its challenges in maintaining innovation and attracting talent [21][22]. - The legal environment in California, which prohibits non-compete agreements, has facilitated the free movement of talent, further accelerating the growth of the "OpenAI Mafia" and the overall innovation landscape in Silicon Valley [24]. Group 4: Capital and Investment Trends - The venture capital landscape in Silicon Valley is heavily concentrated, with over 30% of U.S. venture capital flowing into the region, enabling startups to secure substantial funding even before product launches [25][27]. - The high valuations of AI startups, often exceeding $1 billion without product releases, create a competitive disadvantage for smaller teams, as they struggle to secure resources [27][28]. - The influx of capital has also led to a talent war, with top positions in startups commanding high salaries, reinforcing a cycle where larger companies dominate talent acquisition [28].

Core Viewpoint - Deutsche Telekom is developing an "AI Phone" in collaboration with companies like Perplexity and Picsart, aiming to integrate AI capabilities into a low-cost smartphone targeted at the European market, with a launch planned for the second half of this year and a price under $1000 by 2026 [1][2][4]. Group 1: Product Development - The AI Phone will feature a new AI assistant application called "Magenta AI," which will be available for existing Deutsche Telekom customers on Android and iOS devices [6]. - Perplexity, valued at approximately $9 billion, plays a crucial role in the phone's development, transitioning from a passive answering machine to an active assistant capable of performing tasks like booking flights and sending messages [2][3][7]. - The phone is expected to incorporate AI services from Google Cloud, ElevenLabs, and Picsart, enhancing user experience through proactive assistance [4][6]. Group 2: Market Context - The initiative reflects a broader trend among telecom operators to compete more effectively with tech giants like Apple and Google, who have historically dominated the smartphone market and limited telecom companies' roles [3][4]. - Deutsche Telekom's move into AI-driven hardware represents a strategic effort to strengthen its consumer presence in a highly competitive market, where traditional companies are increasingly viewing AI as a potential solution to regain market relevance [6][7]. - The collaboration with Perplexity marks a significant step for both companies, as they seek to differentiate themselves in a crowded market where even established players like LG have exited [7].

Group 1 - Perplexity is raising a $50 million seed and pre-seed investment fund, with most of the capital coming from limited partners [1] - The company previously raised $500 million at a valuation of $9 billion in December [1] - Perplexity has access to a network of approximately 80,000 developers, allowing it to understand user engagement with its API [1] Group 2 - The fund is managed by general partners Kelly Graziadei and Joanna Lee Shevelenko, who co-founded early-stage venture firm F7 Ventures in 2018 [1] - F7 Ventures has invested in startups such as women's health company Midi [1] - It is unclear whether Graziadei and Shevelenko will continue to operate F7 or focus entirely on Perplexity's venture fund [1]

Core Insights - The valuation multiples for major AI developers like OpenAI and Anthropic have decreased over the past year, while some AI application developers such as Perplexity and Writer have seen their multiples increase [1][2] - Investors are willing to assign high valuations to AI applications due to their revenue growth significantly outpacing that of software applications over the past decade [2] - AI startups targeting specific industries, such as healthcare and legal, are receiving the highest multiples from investors, as they leverage proprietary data to train their models [3] Valuation Trends - Perplexity's annual recurring revenue exceeded $80 million in January, a 60% increase from $50 million in November, leading to a valuation multiple of 170 [2] - Abridge, a six-year-old startup, was valued at $2.5 billion based on its $50 million annual recurring revenue, resulting in a multiple of 50 [4] - Harvey, a three-year-old legal AI startup, was valued at $2.7 billion earlier this year, a decrease from a valuation of $635 million a year ago, despite maintaining a high revenue multiple [5] Revenue Multiples of Major AI Companies - Glean's pre-investment valuation was $4.3 billion, equating to a multiple of 43 based on $100 million annual recurring revenue, which increased to a multiple of 67 with a valuation of $2 billion based on $30 million revenue [6] - Anthropic's valuation is approximately $58 billion based on an annual revenue of $1 billion, down from a multiple of 150 a year ago [6] - OpenAI's latest valuation is $260 billion, corresponding to a multiple of 43 based on an estimated annual revenue of $6 billion, a decrease from a multiple of 54 a year ago [6]

Investment Rating - The report maintains an "Accumulate" rating for the computer industry [6][7]. Core Insights - Perplexity has launched a new version of its Sonar search model, which is optimized for search applications and operates on Cerebras' inference infrastructure. The new model achieves a token decoding speed of 1200 tokens per second, which is over 8.5 times faster than Gemini 2.0 Flash. User evaluations indicate that the performance of the new Sonar model significantly surpasses that of GPT-4omini and Claude 3.5 Haiku, while being comparable to GPT-4o [23][24]. - Huawei's DeepSeek V3/R1 flagship model, with 671 billion parameters, has been fully optimized for deployment on Huawei Cloud's Ascend cloud service, meeting commercial deployment needs [25]. - Tencent's Yuanbao has integrated the DeepSeekR1 model, enabling deep thinking and online search capabilities, enhancing the model's reasoning and response accuracy [26]. - Baidu announced the open-sourcing of its next-generation Wenxin model series 4.5 starting June 30, with a deep search feature available for free from April 1 [27]. Investment Recommendations - The widespread adoption of DeepSeek and improvements in model performance are expected to accelerate the penetration of AI applications. The report is optimistic about the overall industry chain's prosperity and recommends focusing on investment opportunities in the IaaS layer, specifically highlighting companies such as Microsoft (MSFT), Google (GOOGL), Meta (META), and Amazon (AMAZON) [28].