Fed policy, AI and the political regime change have no doubt influenced today’s innovation economy. We explore these trends and more in our recently released State of the Markets report. To broaden these perspectives, we gathered over 100 members of the New York innovation ecosystem for a panel discussion featuring top investor and economist thought leaders.  

  • Phil Neuhart, Director of Market and Economic Research, First Citizens Bank  

Here are highlights and commentary on the themes discussed at the event.  

On DeepSeek’s impact on AI innovation: 

The year 2025 is seen as a pivotal year for reasoning advancements in AI, particularly at the application layer, with companies growing rapidly and efficiently. To date, the rate of AI innovation, especially training LLMs, has been closely correlated with astronomical capital investment and access to hardware. As a result, Nvidia’s revenue has been used as a public market proxy for activity; the AI chipmaker has cornered the market on semiconductors needed to train new models, and its sales are rising in proportion to public adoption of AI.    

However, DeepSeek’s model performance, subject to additional diligence, has downplayed assumptions around the compute arms race narrative. These developments have stirred discussions on Jevon’s paradox – the idea that as cost of systems and models reduces, the resource’s use increases dramatically. Exposing the R1 model as an open-source epiphany should theoretically lower training cost and get to more efficient model reasoning.   

  • “The development of the GPT-4 model, trained across 6-7 trillion parameters, could have been considered one of the most complex artifacts in human history. With DeepSeek, companies and researchers need to understand two vectors of improvement showcasing efficiency in both training data and model reasoning.”  

– George Mathew, Managing Director, Insight Partners  

  • “The DeepSeek release, and the broader promise of continued model efficiency and improved reasoning across models in general, makes it an exciting time for application layer AI where companies that sit on top of these models that can deliver a magical user experience to their customers — enterprise and consumer alike.” 

– Tiffany Luck, Partner, NEA  

On “good” investments in today’s market: 

A flood of capital chasing AI innovation has buoyed valuations, fostering some discussions on a hype-induced bubble. In an evolving market, the benchmarks around how and when companies can raise additional capital are also changing. With revenue growth taking a backseat to efficiency, the typical company raising a Series A today is growing at 69% versus 171% YoY in 2021. Additionally, among companies that raised capital in 2024, the typical Series B company only increased its burn 8% YoY.  

During these times, investors need to be grounded in some uncertainties and truths. Additionally, there are still companies out there with capital structures that need revision.  

  • “Early internet is still the best analogy for AI. With uncertainties about who will be the winners, Founder Collective is focusing on what has always worked – customer obsessed founders building real businesses that delight users.” 
– Amanda Herson, General Partner, Founder Collective   
  • “NEA has maintained the same process, regardless of sector: (1) the why – a burning problem founders are drawn to working on 24/7; (2) growth at scale and underlying fundamentals is the second pre-requisite.; (3) there is limited time / capital to try things, so what I call problem-solution matching with real buyer intent is the third piece; and (4) efficient repeatability of the first three pieces.”  

– Tiffany Luck, Partner, NEA  

  • “Good companies need capital and need to be able to scale; valuation is not indicative of the real value / long term outcome of the business. Down rounds are not 4-letter words; some of the best portfolio companies have down rounds and are perfectly OK.” 

– George Mathew, Managing Director, Insight Partners  

On investing outside and alongside AI: 

Although AI is reshaping entire value chains and business models, there are certain sectors ripe for innovation. Unsurprisingly, enterprise software and frontier tech focused companies seeing the biggest boom from AI innovations have captured disproportionate investment for their capital expenditures.   

Additionally, autonomous vehicles and defense tech are also emerging as growth areas. Autonomous vehicles, buoyed by every more powerful AI models, are driving a large share of the investment in frontier tech, which has jumped from the fourth-most heavily invested sector in 2022 to the second-most favored sector in 2024. Defense technology is also emerging as a growth area for frontier tech investors, with notable deals for several defense tech unicorns such as Anduril among the largest deals of the year.  

Tech investors’ presence at President Trump’s inauguration has drawn some cautious optimism on efficiency driven by deregulation and government change. This could have disproportionate benefits in financial services and healthcare.  

  • “Wins are found in pockets people aren’t thinking about but a founder has a unique point of view in.”  

– Amanda Herson, General Partner, Founder Collective  

  • “Company building itself is changing. The recipe for new software isn’t just sprinkling AI pixie dust. We need to find founders looking to upend the entire value chain, creating complimentary synthetic systems alongside the humans that do the work.”  

– George Mathew, Managing Director, Insight Partners  

  • “Predicting that regulatory changes will feed into M&A pickup as early as Q2 with a lot of pent-up demand slated into this year. Healthcare is low hanging fruit, financial can find a swing into fintech. Building nuclear could be easier in 6 months. However, we can’t get too carried away on the tailwinds only because there are still risks. Tariffs and immigration reform could upend any expected positive dynamics.”  

– Phil Neuhart, Director of Market and Economic Research, First Citizens Bank

On public market readiness and IPOs: 

Expectations are that the 2nd half the year will see a pickup in new public market issuances, going out through 2026. Secondary markets and M&A activity may provide some liquidity to unicorns, but IPOs will need to play a key role. We estimate the low bar for IPO readiness includes over $100M run rate for revenue, at least 15% YoY growth, and greater than negative 25% margin. Currently, only about only one quarter of the unicorn cohort are IPO hopefuls.  

Companies testing public markets prior to that are likely testing a higher bar, as unicorn companies are determining how their KPIs stack up against investor expectations. Being able to prove that accurate 12-month projections is a pre-requisite for these companies.  

  • “Public markets want pure plays for technological innovation, for which IPOs are necessary. Investors want and need fresh places to put capital. Last year’s narrative was waiting for election results to wait on IPOs; IPOs don’t happen in 4-year cycles but this regime change marked a change in readiness and motivation. With concerns around AI hype, this market looks like 2004 not 1999 – fundamentals, path to profitability all matter more. We are unlikely to go back 20 years to relearn fresh lessons in public markets.” 

– Phil Neuhart, Director of Market and Economic Research, First Citizens Bank  

  

To learn more about these trends and understand what they can mean for your business, read the full 2025 State of the Markets H1 2025 report, or contact Ash Bhatia (abhatia@svb.com).