Venture Bytes #114: Top Three Tech Trends For 2025

Top Three Tech Trends For 2025

Despite a decline in US venture capital investment from previous peaks, capital inflowsi nto US startups remain strong, exceeding levels seen in 26 of the last 30 years, according to Silicon Valley Bank. Consequently, we have identified three key technology trends to monitor in 2025, focusing on the technologies poised to reach a critical inflection point during the year:
● Technologies Boosting Performance & Efficiency of Data Centers and Chips:
Best Startup Idea – Lightmatter
● Global Embrace of AI Agents: Best Startup Idea – Sierra
● Widespread Adoption of Humanoid Robots: Best Startup Idea – Figure (Disclosure:
Figure is a Portfolio Company of MVP All-Star Fund V)

Increased Focus on Technologies Boosting Performance & Efficiency of Data Centers and Chips

‍
The growing demand for data centers drives significant environmental costs, particularly electricity, water, and carbon emissions. US data centers consumed 4% of the country’s electricity in 2022, with projections suggesting this will rise to 6% by 2026, per the International Energy Agency. This increase is largely driven by the higher power consumption of AI queries — ChatGPT uses 2.9 Wh per request, compared to Google’s 0.3 Wh. As AI adoption grows, data centers may require an additional 160- 590 TWh by 2026, equivalent to the total energy consumption of countries like Sweden or Germany.

‍
Data centers contribute around 2% of total greenhouse gas emissions, comparable to the entire airline industry, per Anthesis. Water usage is another concern, with US data centers consuming 1.7 billion liters daily for cooling, per Data Center Knowledge. Moreover, data centers contribute significantly to carbon emissions. The IT sector, including data centers, is responsible for 2.5-3.7% of global carbon emissions, per The Shift Project, higher than the 2.4% of the aviation industry, per EESI. Moreover, these official numbers are artificially deflated by renewab energy credits, underscoring that the problem is much bigger.

‍

Photonic technology startups are tackling such issues, by reducing heat generation, enabling faster data transmission, and enhancing resilience to environmental factors, ultimately improving AI training performance. Lightmatter, a California- based startup, utilizes photonic technology to improve the efficiency of data centers; its upcoming product Passage, set for release by 2026, is designed to integrate over 1 million GPUs in parallel during AI training. If Lightmatter chips replaced Nvidia DGX-A100 processors in 100 data centers, the energy savings would equal the annual electricity consumption of 61.6 million homes, per Aim Research.

Global Embrace of AI Agents

The potential for AI agents to deliver measurable benefits is rapidly gaining attention across industries. According to McKinsey & Co., AI has already demonstrated significant improvements in operational efficiency, including a 14% increase in customer support resolution rates per hour and a 9% reduction in handling times. These performance gains translate to productivity improvements of 30-45% in various sectors, underscoring AI’s capacity to drive cost savings and operational excellence. Additionally, AI agents have proven to expedite content review cycles by 20-60%, further highlighting their efficiency.

As these benefits become more evident, the adoption of AI agents is accelerating. A recent Capgemini survey found that 82% of companies plan to integrate AI agents within the next three years, with 48% already using them and an additional 33% actively exploring options, according to Forum Ventures. This rapid adoption is reflective of a broader trend in which AI is expected to disrupt a wide range of industries by automating critical tasks and roles.

The opportunity for AI agents is vast. A16z and the Bureau of Labor Statistics estimate that over the next decade, AI could automate as many as 50 major job roles, encompassing professions such as architects, financial advisors, lawyers, and software developers. This represents about one-third of the US workforce, underscoring the transformative potential of AI across sectors. As businesses continue to embrace AI solutions, the market for AI agents is poised for explosive growth, creating new value propositions and competitive advantages for early adopters.

Sierra, a California-based startup, offers personalized text and voice AI agents. Tangible benefits to customers affirm the credibility of the startup. For instance, Casper, a customer in the mattress business, boosted customer engagement by 20% while maintaining a 74% resolution rate. WeightWatchers, another customer in health tech, achieved a CSAT of 4.6/5 with a resolution rate of 70%. A valuation markup of 458% in the latest funding round affirms investor confidence. Marquee investors include Benchmark, ICONIQ Capital, and Sequoia Capital, to name a few.

Widespread Adoption of Humanoid Robots

The  robotics  industry  is  rapidly  evolving,  fueled  by  AI advancements. Humanoid robots are attracting significant investor interest across sectors like manufacturing and healthcare, among others. In 2023, the sector received $292.7 million in investments, with 2024 surpassing that amount, per Pitchbook. Notable developments include the Hyundai-owned Boston Dynamics’ Atlas robot, which recently showcased its ability to perform complex tasks in a simulated factory, indicating progress for industrial applications. Tesla also demonstrated its humanoid robot, Optimus, at the “We, Robot” event.

Humanoid robots are poised to disrupt traditional workforce models by offering unmatched versatility, 24/7 productivity, and injury-free performance. With labor shortages and rising workplace injuries—US operational injuries increased from 2.1 million in 2020 to 2.4 million in 2023, and fatal work injuries reached 5,486 in 2022 (US Department of Labor)—industries face growing challenges. Additionally, 75% of employers struggled to fill job vacancies in critical sectors, with 7 million unfilled positions in 2023, per Manpower Group. The aging population and declining birth rates exacerbate these issues.

Humanoid robots, powered by advances in AI, machine learning, and computer vision, offer a solution. The humanoid robot market is projected to grow from $2 billion in 2024 to $13.3 billion by 2029, at a CAGR of 45.5% per MarketsandMarkets. Startups like Figure are capitalizing on this growth. In February 2024, Figure, a California-based humanoid robot startup, raised $675 million in Series B funding, boosting its valuation to $2.7 billion—a 448% increase from July 2023. Investors include Intel Capital, M12, Bezos Expeditions, and OpenAI, among others.

Within 31 months of incorporation, Figure is shipping its second model, Figure 02, to commercial clients, as announced by CEO Brett Adcock. Figure joins Agility Robotics’ Digit as one of the few humanoid robotics companies deploying robots in paying clients’ warehouses and factories. In January, a tethered Figure 01 moved at only 17% of a human’s speed. Figure 02, however, is seven times faster, according to the company. Although humans remain faster, a 7X speed improveme       a year signals rapid progress in humanoid robotics.

‍

Future of Battlefield is AI and US Needs to Stay Ahead

The pace of change in wars has never been faster, with technology compressing timelines and reshaping the nature of engagement. These accelerated cycles demand an unprecedented level of adaptability. In 2024, with armed conflicts at levels not seen since World War II, defense organizations face a perfect storm of complexity and urgency.

The US defense industry finds itself at a critical juncture, where its current business model is increasingly unsustainable in the face of asymmetric threats and evolving technologies. For example, intercepting low-cost Houthi drones—built for a few thousand dollars—required US naval missiles costing up to $2.1 million per shot. This imbalance highlights the urgent need for a technological shift. At the center of this transformation is artificial intelligence (AI), which is emerging as the most transformative force in defense technology, poised to redefine how nations safeguard their interests and respond to evolving threats.

The air-to-air drone conflict in Ukraine exemplifies war’s rapid evolution. What began with Ukrainian forces deploying first- person-view drones to ram Russian drones quickly escalated into a weekly cycle of adaptation: Russian operators countered with new tactics, prompting Ukrainians to arm drones with munitions and develop AI targeting systems. In turn, Russians experimented with dazzle paint schemes to evade detection. This constant back-and-forth highlights the urgent need for nations like the US to embrace and lead in AI-driven defense advancements.

While the size of armed forces and relative military spending have declined in many countries, overall defense budgets and technological sophistication continue to grow. According to the Global Peace Index, per capita military spending and the use of advanced systems have surged over the past three decades, reflecting a focus on fewer but more capable forces equipped with cutting-edge technology.

This shift has made defense technology a hotbed for venture capital investment. By mid-November, defense tech startups had raised approximately $3 billion across 85 funding rounds, surpassing the total capital raised in 2021 – a record year for venture funding.

At the 2024 DefenseTech Summit in Tel Aviv, Sequoia Capital partner Shaun Maguire highlighted defense tech as one of the most attractive sectors for investors. This trend is supported by structural changes within the Department of Defense (DoD), including initiatives like the Defense Innovation Unit, AFWERX, and DARPA’s increased collaboration with startups.

Companies and investors who were once reluctant to engage with the defense sector are now shifting their perspectives. OpenAI recently joined the ranks of major AI players, such as Meta and Anthropic, by revealing its collaboration with the defense industry. In partnership with Anduril, OpenAI will provide its models to enhance air defense systems. Meanwhile, the landscape of specialized defense and dual-use investors has seen dramatic expansion, growing from a mere handful in 2022 to over 100 today, reflecting a broader acceptance and recognition of the sector’s strategic importance.

AI is central to these developments, with the DoD’s unclassified investments in AI growing from $600 million in FY16 to $1.8 billion in 2024, encompassing over 685 active projects. NATO has also increased its defense investments by 11% in real terms, while European funds are mobilizing significant capital for defense tech.

This influx of resources is fueling groundbreaking work in areas like AI-driven decision-making. For instance, Project Maven in the US enhances intelligence analysis and targeting, while Ukraine’s Kropyva and GIS Arta systems provide real-time battlefield intelligence for artillery. Similarly, Israel’s AI platforms, such as Gospel          le rapid target identification in conflict zones.

‍

California-based Shield AI is revolutionizing military aviation. Its main product, called Hivemind, enables intelligent aircraft to operate autonomously in high-threat environments on the edge, the company said. This operation is without the need of GPS, remote operators or command inputs. Shield AI’s state-of-the- art autonomous pilot systems and drones are actively deployed in Ukraine’s combat zones, highlighting the transformative impact of advanced AI and drone technology in modern warfare. These innovations provide critical tactical advantages, enabling effective operations in highly dynamic and demanding battlefield conditions.

Yet,  significant  challenges  remain.  Autonomous  systems are integral to military operations worldwide, but the lack of infrastructure to unify these technologies limits their effectiveness. Operators often manage these systems in isolation, reducing their efficiency. Companies like Picogrid are stepping in to address this gap, creating integrated ecosystems that connect sensors, cameras, and autonomous platforms, delivering cohesive capabilities at scale. Picogrid has secured more than a dozen federal contracts and successfully deployed over 40 hardware systems across six U.S. military bases, with additional installations operating in Ukraine.

AI can also play a crucial role in reducing civilian casualties by enhancing precautionary measures in conflict zones. Through real-time mapping of civilian infrastructure and continuous risk assessments, AI can provide military forces with detailed insights to identify and avoid civilian areas during operations. By analyzing evolving conditions and potential threats, AI enables more precise targeting, ensuring that military action takke every possible step to minimize collateral damage.

‍

‍